ylhan/note_sync
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
14c280ffee1388c74ef8347008d464d56e377418
note_sync/大模型/deepseek/_scripts/raw_deepseek.json
T

3203 lines
206 KiB
JSON
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
[
{
"id": "2502.10299v1",
"version_id": "http://arxiv.org/abs/2502.10299v1",
"title": "Open-Source AI-Powered Optimization in Scalene: Advancing Python Performance Profiling with DeepSeek-R1 and LLaMA 3.2",
"published": "2025-02-14T17:01:06Z",
"updated": "2025-02-14T17:01:06Z",
"authors": [
"Saem Hasan",
"Sanju Basak"
],
"summary": "Python's flexibility and ease of use come at the cost of performance inefficiencies, requiring developers to rely on profilers to optimize execution. SCALENE, a high-performance CPU, GPU, and memory profiler, provides fine-grained insights into Python applications while running significantly faster than traditional profilers. Originally, SCALENE integrated OpenAI's API to generate AI-powered optimization suggestions, but its reliance on a proprietary API limited accessibility. This study explores the feasibility of using opensource large language models (LLMs), such as DeepSeek-R1 and Llama 3.2, to generate optimization recommendations within SCALENE. Our evaluation reveals that DeepSeek-R1 provides effective code optimizations comparable to proprietary models. We integrate DeepSeek-R1 into SCALENE to automatically analyze performance bottlenecks and suggest improvements, enhancing SCALENE's utility while maintaining its open-source nature. This study demonstrates that open-source LLMs can be viable alternatives for AI-driven code optimization, paving the way for more accessible and cost-effective performance analysis tools.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.PL"
},
{
"id": "2502.11096v1",
"version_id": "http://arxiv.org/abs/2502.11096v1",
"title": "Mixture of Tunable Experts -- Behavior Modification of DeepSeek-R1 at Inference Time",
"published": "2025-02-16T12:24:39Z",
"updated": "2025-02-16T12:24:39Z",
"authors": [
"Robert Dahlke",
"Henrik Klagges",
"Dan Zecha",
"Benjamin Merkel",
"Sven Rohr",
"Fabian Klemm"
],
"summary": "We present the Mixture-of-Tunable-Experts (MoTE), a method that extends the Mixture-of-Experts architecture of Large Language Models (LLMs). Without additional training, MoTE enables meaningful and focused behavior changes in LLMs on-the-fly during inference time. By analyzing the digital LLM brain of DeepSeek-R1 using a technique we dub 'functional Token Resonance Imaging' (fTRI) -- inspired by fMRI and using prompts designed to elicit specific behavior (e.g., 'What happened {time}{place}?') -- we empirically identify distinctive experts associated with behaviors like refusal responses. Using MoTE we are able to intervene and control such specific behavior. We switched off the top 10 most refusal-relevant experts (0.07% of R1's 14,848 routed experts), achieving a 52% refusal reduction on sensitive reference prompts without performance degradation on MT-Bench. Random expert deactivation resulted in smaller behavioral shifts with increased noise, whereas forced expert activation led to significantly higher refusal rates. Our approach shares similarities with sparse autoencoders (SAEs) in terms of explainability and steerability. Unlike SAEs, MoTE does not require large training efforts, as within MoEs with a vast number of experts, specialization already emerged naturally during pretraining. Our findings suggest that significant functional mechanisms in Mixture-of-Experts architectures can at least partially be localized in a small number of specific experts, rather than being distributed throughout the model's weights. Expert subgroups can be tuned to trigger significant behavior variations, providing insights into the inner workings of LLMs.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2502.11108v1",
"version_id": "http://arxiv.org/abs/2502.11108v1",
"title": "Knowledge Graph-Driven Retrieval-Augmented Generation: Integrating Deepseek-R1 with Weaviate for Advanced Chatbot Applications",
"published": "2025-02-16T12:52:28Z",
"updated": "2025-02-16T12:52:28Z",
"authors": [
"Alexandru Lecu",
"Adrian Groza",
"Lezan Hawizy"
],
"summary": "Large language models (LLMs) have significantly advanced the field of natural language generation. However, they frequently generate unverified outputs, which compromises their reliability in critical applications. In this study, we propose an innovative framework that combines structured biomedical knowledge with LLMs through a retrieval-augmented generation technique. Our system develops a thorough knowledge graph by identifying and refining causal relationships and named entities from medical abstracts related to age-related macular degeneration (AMD). Using a vector-based retrieval process and a locally deployed language model, our framework produces responses that are both contextually relevant and verifiable, with direct references to clinical evidence. Experimental results show that this method notably decreases hallucinations, enhances factual precision, and improves the clarity of generated responses, providing a robust solution for advanced biomedical chatbot applications.",
"comment": "",
"journal_ref": "",
"doi": "10.1016/j.procs.2024.10.219",
"primary": "cs.CL"
},
{
"id": "2502.11137v3",
"version_id": "http://arxiv.org/abs/2502.11137v3",
"title": "Safety Evaluation of DeepSeek Models in Chinese Contexts",
"published": "2025-02-16T14:05:54Z",
"updated": "2025-05-08T01:35:18Z",
"authors": [
"Wenjing Zhang",
"Xuejiao Lei",
"Zhaoxiang Liu",
"Ning Wang",
"Zhenhong Long",
"Peijun Yang",
"Jiaojiao Zhao",
"Minjie Hua",
"Chaoyang Ma",
"Kai Wang",
"Shiguo Lian"
],
"summary": "Recently, the DeepSeek series of models, leveraging their exceptional reasoning capabilities and open-source strategy, is reshaping the global AI landscape. Despite these advantages, they exhibit significant safety deficiencies. Research conducted by Robust Intelligence, a subsidiary of Cisco, in collaboration with the University of Pennsylvania, revealed that DeepSeek-R1 has a 100\\% attack success rate when processing harmful prompts. Additionally, multiple safety companies and research institutions have confirmed critical safety vulnerabilities in this model. As models demonstrating robust performance in Chinese and English, DeepSeek models require equally crucial safety assessments in both language contexts. However, current research has predominantly focused on safety evaluations in English environments, leaving a gap in comprehensive assessments of their safety performance in Chinese contexts. In response to this gap, this study introduces CHiSafetyBench, a Chinese-specific safety evaluation benchmark. This benchmark systematically evaluates the safety of DeepSeek-R1 and DeepSeek-V3 in Chinese contexts, revealing their performance across safety categories. The experimental results quantify the deficiencies of these two models in Chinese contexts, providing key insights for subsequent improvements. It should be noted that, despite our efforts to establish a comprehensive, objective, and authoritative evaluation benchmark, the selection of test samples, characteristics of data distribution, and the setting of evaluation criteria may inevitably introduce certain biases into the evaluation results. We will continuously optimize the evaluation benchmark and periodically update this report to provide more comprehensive and accurate assessment outcomes. Please refer to the latest version of the paper for the most recent evaluation results and conclusions.",
"comment": "12 pages, 2 tables, 7 figures",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2502.11164v5",
"version_id": "http://arxiv.org/abs/2502.11164v5",
"title": "Quantifying the Capability Boundary of DeepSeek Models: An Application-Driven Performance Analysis",
"published": "2025-02-16T15:29:58Z",
"updated": "2025-05-16T01:20:01Z",
"authors": [
"Kaikai Zhao",
"Zhaoxiang Liu",
"Xuejiao Lei",
"Jiaojiao Zhao",
"Zhenhong Long",
"Zipeng Wang",
"Ning Wang",
"Meijuan An",
"Qingliang Meng",
"Peijun Yang",
"Minjie Hua",
"Chaoyang Ma",
"Wen Liu",
"Kai Wang",
"Shiguo Lian"
],
"summary": "DeepSeek-R1, known for its low training cost and exceptional reasoning capabilities, has achieved state-of-the-art performance on various benchmarks. However, detailed evaluations for DeepSeek Series models from the perspective of real-world applications are lacking, making it challenging for users to select the most suitable DeepSeek models for their specific needs. To address this gap, we presents the first comprehensive evaluation of the DeepSeek and its related models (including DeepSeek-V3, DeepSeek-R1, DeepSeek-R1-Distill-Qwen series, DeepSeek-R1-Distill-Llama series, their corresponding 4-bit quantized models, and the reasoning model QwQ-32B) using our enhanced A-Eval benchmark, A-Eval-2.0. Our systematic analysis reveals several key insights: (1) Given identical model architectures and training data, larger parameter models demonstrate superior performance, aligning with the scaling law. However, smaller models may achieve enhanced capabilities when employing optimized training strategies and higher-quality data; (2) Reasoning-enhanced model show significant performance gains in logical reasoning tasks but may underperform in text understanding and generation tasks; (3) As the data difficulty increases, distillation or reasoning enhancements yield higher performance gains for the models. Interestingly, reasoning enhancements can even have a negative impact on simpler problems; (4) Quantization impacts different capabilities unevenly, with significant drop on logical reasoning and minimal impact on text generation. Based on these results and findings, we design an model selection handbook enabling users to select the most cost-effective models without efforts.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2502.11347v1",
"version_id": "http://arxiv.org/abs/2502.11347v1",
"title": "Evaluating the Performance of the DeepSeek Model in Confidential Computing Environment",
"published": "2025-02-17T01:55:38Z",
"updated": "2025-02-17T01:55:38Z",
"authors": [
"Ben Dong",
"Qian Wang"
],
"summary": "The increasing adoption of Large Language Models (LLMs) in cloud environments raises critical security concerns, particularly regarding model confidentiality and data privacy. Confidential computing, enabled by Trusted Execution Environments (TEEs), offers a promising solution to mitigate these risks. However, existing TEE implementations, primarily CPU-based, struggle to efficiently support the resource-intensive nature of LLM inference and training. In this work, we present the first evaluation of the DeepSeek model within a TEE-enabled confidential computing environment, specifically utilizing Intel Trust Domain Extensions (TDX). Our study benchmarks DeepSeek's performance across CPU-only, CPU-GPU hybrid, and TEE-based implementations. For smaller parameter sets, such as DeepSeek-R1-1.5B, the TDX implementation outperforms the CPU version in executing computations within a secure environment. It highlights the potential for efficiently deploying LLM models on resource-constrained systems while ensuring security. The overall GPU-to-CPU performance ratio averages 12 across different model sizes, with smaller models exhibiting a lower ratio. Additionally, we provide foundational insights and guidance on optimizing CPU-GPU confidential computing solutions for scalable and secure AI deployments. Our findings contribute to the advancement of privacy-preserving AI, paving the way for efficient and secure LLM inference in confidential computing environments.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.PF"
},
{
"id": "2502.12825v2",
"version_id": "http://arxiv.org/abs/2502.12825v2",
"title": "Reasoning and the Trusting Behavior of DeepSeek and GPT: An Experiment Revealing Hidden Fault Lines in Large Language Models",
"published": "2025-02-18T12:46:18Z",
"updated": "2025-02-19T11:57:19Z",
"authors": [
"Rubing Li",
"João Sedoc",
"Arun Sundararajan"
],
"summary": "When encountering increasingly frequent performance improvements or cost reductions from a new large language model (LLM), developers of applications leveraging LLMs must decide whether to take advantage of these improvements or stay with older tried-and-tested models. Low perceived switching frictions can lead to choices that do not consider more subtle behavior changes that the transition may induce. Our experiments use a popular game-theoretic behavioral economics model of trust to show stark differences in the trusting behavior of OpenAI's and DeepSeek's models. We highlight a collapse in the economic trust behavior of the o1-mini and o3-mini models as they reconcile profit-maximizing and risk-seeking with future returns from trust, and contrast it with DeepSeek's more sophisticated and profitable trusting behavior that stems from an ability to incorporate deeper concepts like forward planning and theory-of-mind. As LLMs form the basis for high-stakes commercial systems, our results highlight the perils of relying on LLM performance benchmarks that are too narrowly defined and suggest that careful analysis of their hidden fault lines should be part of any organization's AI strategy.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2502.12893v2",
"version_id": "http://arxiv.org/abs/2502.12893v2",
"title": "H-CoT: Hijacking the Chain-of-Thought Safety Reasoning Mechanism to Jailbreak Large Reasoning Models, Including OpenAI o1/o3, DeepSeek-R1, and Gemini 2.0 Flash Thinking",
"published": "2025-02-18T14:29:12Z",
"updated": "2025-02-27T01:32:58Z",
"authors": [
"Martin Kuo",
"Jianyi Zhang",
"Aolin Ding",
"Qinsi Wang",
"Louis DiValentin",
"Yujia Bao",
"Wei Wei",
"Hai Li",
"Yiran Chen"
],
"summary": "Large Reasoning Models (LRMs) have recently extended their powerful reasoning capabilities to safety checks-using chain-of-thought reasoning to decide whether a request should be answered. While this new approach offers a promising route for balancing model utility and safety, its robustness remains underexplored. To address this gap, we introduce Malicious-Educator, a benchmark that disguises extremely dangerous or malicious requests beneath seemingly legitimate educational prompts. Our experiments reveal severe security flaws in popular commercial-grade LRMs, including OpenAI o1/o3, DeepSeek-R1, and Gemini 2.0 Flash Thinking. For instance, although OpenAI's o1 model initially maintains a high refusal rate of about 98%, subsequent model updates significantly compromise its safety; and attackers can easily extract criminal strategies from DeepSeek-R1 and Gemini 2.0 Flash Thinking without any additional tricks. To further highlight these vulnerabilities, we propose Hijacking Chain-of-Thought (H-CoT), a universal and transferable attack method that leverages the model's own displayed intermediate reasoning to jailbreak its safety reasoning mechanism. Under H-CoT, refusal rates sharply decline-dropping from 98% to below 2%-and, in some instances, even transform initially cautious tones into ones that are willing to provide harmful content. We hope these findings underscore the urgent need for more robust safety mechanisms to preserve the benefits of advanced reasoning capabilities without compromising ethical standards.",
"comment": "Website: https://maliciouseducator.org/",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2502.17487v2",
"version_id": "http://arxiv.org/abs/2502.17487v2",
"title": "User Intent to Use DeepSeek for Healthcare Purposes and their Trust in the Large Language Model: Multinational Survey Study",
"published": "2025-02-18T16:49:36Z",
"updated": "2025-03-02T06:05:40Z",
"authors": [
"Avishek Choudhury",
"Yeganeh Shahsavar",
"Hamid Shamszare"
],
"summary": "Large language models (LLMs) increasingly serve as interactive healthcare resources, yet user acceptance remains underexplored. This study examines how ease of use, perceived usefulness, trust, and risk perception interact to shape intentions to adopt DeepSeek, an emerging LLM-based platform, for healthcare purposes. A cross-sectional survey of 556 participants from India, the United Kingdom, and the United States was conducted to measure perceptions and usage patterns. Structural equation modeling assessed both direct and indirect effects, including potential quadratic relationships. Results revealed that trust plays a pivotal mediating role: ease of use exerts a significant indirect effect on usage intentions through trust, while perceived usefulness contributes to both trust development and direct adoption. By contrast, risk perception negatively affects usage intent, emphasizing the importance of robust data governance and transparency. Notably, significant non-linear paths were observed for ease of use and risk, indicating threshold or plateau effects. The measurement model demonstrated strong reliability and validity, supported by high composite reliabilities, average variance extracted, and discriminant validity measures. These findings extend technology acceptance and health informatics research by illuminating the multifaceted nature of user adoption in sensitive domains. Stakeholders should invest in trust-building strategies, user-centric design, and risk mitigation measures to encourage sustained and safe uptake of LLMs in healthcare. Future work can employ longitudinal designs or examine culture-specific variables to further clarify how user perceptions evolve over time and across different regulatory environments. Such insights are critical for harnessing AI to enhance outcomes.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2502.14926v3",
"version_id": "http://arxiv.org/abs/2502.14926v3",
"title": "DeepSeek-V3, GPT-4, Phi-4, and LLaMA-3.3 generate correct code for LoRaWAN-related engineering tasks",
"published": "2025-02-19T23:16:29Z",
"updated": "2025-04-07T18:44:32Z",
"authors": [
"Daniel Fernandes",
"João P. Matos-Carvalho",
"Carlos M. Fernandes",
"Nuno Fachada"
],
"summary": "This paper investigates the performance of 16 Large Language Models (LLMs) in automating LoRaWAN-related engineering tasks involving optimal placement of drones and received power calculation under progressively complex zero-shot, natural language prompts. The primary research question is whether lightweight, locally executed LLMs can generate correct Python code for these tasks. To assess this, we compared locally run models against state-of-the-art alternatives, such as GPT-4 and DeepSeek-V3, which served as reference points. By extracting and executing the Python functions generated by each model, we evaluated their outputs on a zero-to-five scale. Results show that while DeepSeek-V3 and GPT-4 consistently provided accurate solutions, certain smaller models -- particularly Phi-4 and LLaMA-3.3 -- also demonstrated strong performance, underscoring the viability of lightweight alternatives. Other models exhibited errors stemming from incomplete understanding or syntactic issues. These findings illustrate the potential of LLM-based approaches for specialized engineering applications while highlighting the need for careful model selection, rigorous prompt design, and targeted domain fine-tuning to achieve reliable outcomes.",
"comment": "The peer-reviewed version of this paper is published in Electronics at https://doi.org/10.3390/electronics14071428. This version is typeset by the authors and differs only in pagination and typographical detail",
"journal_ref": "Electronics. 2025; 14(7):1428",
"doi": "10.3390/electronics14071428",
"primary": "cs.SE"
},
{
"id": "2502.14837v2",
"version_id": "http://arxiv.org/abs/2502.14837v2",
"title": "Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs",
"published": "2025-02-20T18:50:42Z",
"updated": "2025-10-03T15:37:19Z",
"authors": [
"Tao Ji",
"Bin Guo",
"Yuanbin Wu",
"Qipeng Guo",
"Lixing Shen",
"Zhan Chen",
"Xipeng Qiu",
"Qi Zhang",
"Tao Gui"
],
"summary": "Multi-head Latent Attention (MLA) is an innovative architecture proposed by DeepSeek, designed to ensure efficient and economical inference by significantly compressing the Key-Value (KV) cache into a latent vector. Compared to MLA, standard LLMs employing Multi-Head Attention (MHA) and its variants such as Grouped-Query Attention (GQA) exhibit significant cost disadvantages. Enabling well-trained LLMs (e.g., Llama) to rapidly adapt to MLA without pre-training from scratch is both meaningful and challenging. This paper proposes the first data-efficient fine-tuning method for transitioning from MHA to MLA (MHA2MLA), which includes two key components: for partial-RoPE, we remove RoPE from dimensions of queries and keys that contribute less to the attention scores, for low-rank approximation, we introduce joint SVD approximations based on the pre-trained parameters of keys and values. These carefully designed strategies enable MHA2MLA to recover performance using only a small fraction (0.3% to 0.6%) of the data, significantly reducing inference costs while seamlessly integrating with compression techniques such as KV cache quantization. For example, the KV cache size of Llama2-7B is reduced by 92.19%, with only a 0.5% drop in LongBench performance.",
"comment": "16 pages, 8 figures; Accepted to ACL 2025",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2502.18505v1",
"version_id": "http://arxiv.org/abs/2502.18505v1",
"title": "Comprehensive Analysis of Transparency and Accessibility of ChatGPT, DeepSeek, And other SoTA Large Language Models",
"published": "2025-02-21T23:53:13Z",
"updated": "2025-02-21T23:53:13Z",
"authors": [
"Ranjan Sapkota",
"Shaina Raza",
"Manoj Karkee"
],
"summary": "Despite increasing discussions on open-source Artificial Intelligence (AI), existing research lacks a discussion on the transparency and accessibility of state-of-the-art (SoTA) Large Language Models (LLMs). The Open Source Initiative (OSI) has recently released its first formal definition of open-source software. This definition, when combined with standard dictionary definitions and the sparse published literature, provide an initial framework to support broader accessibility to AI models such as LLMs, but more work is essential to capture the unique dynamics of openness in AI. In addition, concerns about open-washing, where models claim openness but lack full transparency, has been raised, which limits the reproducibility, bias mitigation, and domain adaptation of these models. In this context, our study critically analyzes SoTA LLMs from the last five years, including ChatGPT, DeepSeek, LLaMA, and others, to assess their adherence to transparency standards and the implications of partial openness. Specifically, we examine transparency and accessibility from two perspectives: open-source vs. open-weight models. Our findings reveal that while some models are labeled as open-source, this does not necessarily mean they are fully open-sourced. Even in the best cases, open-source models often do not report model training data, and code as well as key metrics, such as weight accessibility, and carbon emissions. To the best of our knowledge, this is the first study that systematically examines the transparency and accessibility of over 100 different SoTA LLMs through the dual lens of open-source and open-weight models. The findings open avenues for further research and call for responsible and sustainable AI practices to ensure greater transparency, accountability, and ethical deployment of these models.(DeepSeek transparency, ChatGPT accessibility, open source, DeepSeek open source)",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.SE"
},
{
"id": "2502.18513v1",
"version_id": "http://arxiv.org/abs/2502.18513v1",
"title": "Analyzing User Perceptions of Large Language Models (LLMs) on Reddit: Sentiment and Topic Modeling of ChatGPT and DeepSeek Discussions",
"published": "2025-02-22T17:00:42Z",
"updated": "2025-02-22T17:00:42Z",
"authors": [
"Krishnaveni Katta"
],
"summary": "While there is an increased discourse on large language models (LLMs) like ChatGPT and DeepSeek, there is no comprehensive understanding of how users of online platforms, like Reddit, perceive these models. This is an important omission because public opinion can influence AI development, trust, and future policy. This study aims at analyzing Reddit discussions about ChatGPT and DeepSeek using sentiment and topic modeling to advance the understanding of user attitudes. Some of the significant topics such as trust in AI, user expectations, potential uses of the tools, reservations about AI biases, and ethical implications of their use are explored in this study. By examining these concerns, the study provides a sense of how public sentiment might shape the direction of AI development going forward. The report also mentions whether users have faith in the technology and what they see as its future. A word frequency approach is used to identify broad topics and sentiment trends. Also, topic modeling through the Latent Dirichlet Allocation (LDA) method identifies top topics in users' language, for example, potential benefits of LLMs, their technological applications, and their overall social ramifications. The study aims to inform developers and policymakers by making it easier to see how users comprehend and experience these game-changing technologies.",
"comment": "13 pages, 8 figures",
"journal_ref": "",
"doi": "",
"primary": "cs.SI"
},
{
"id": "2502.16428v1",
"version_id": "http://arxiv.org/abs/2502.16428v1",
"title": "Visual Reasoning Evaluation of Grok, Deepseek Janus, Gemini, Qwen, Mistral, and ChatGPT",
"published": "2025-02-23T04:01:43Z",
"updated": "2025-02-23T04:01:43Z",
"authors": [
"Nidhal Jegham",
"Marwan Abdelatti",
"Abdeltawab Hendawi"
],
"summary": "Traditional evaluations of multimodal large language models (LLMs) have been limited by their focus on single-image reasoning, failing to assess crucial aspects like contextual understanding, reasoning stability, and uncertainty calibration. This study addresses these limitations by introducing a novel benchmark that integrates multi-image reasoning tasks with rejection-based evaluation and positional bias detection. To evaluate these dimensions, we further introduce entropy as a novel metric for quantifying reasoning consistency across reordered answer variants. We applied this benchmark to assess Grok 3, ChatGPT-4o, ChatGPT-o1, Gemini 2.0 Flash Experimental, DeepSeek Janus models, Qwen2.5-VL-72B-Instruct, QVQ-72B-Preview, and Pixtral 12B across eight visual reasoning tasks, including difference spotting and diagram interpretation. Our findings reveal ChatGPT-o1 leading in overall accuracy (82.5\\%) and rejection accuracy (70.0\\%), closely followed by Gemini 2.0 Flash Experimental (70.8\\%). QVQ-72B-Preview demonstrated superior rejection accuracy (85.5\\%). Notably, Pixtral 12B (51.7\\%) showed promise in specific domains, while Janus models exhibited challenges in bias and uncertainty calibration, reflected in low rejection accuracies and high entropy scores. High entropy scores in Janus models (Janus 7B: 0.8392, Janus 1B: 0.787) underscore their susceptibility to positional bias and unstable reasoning, contrasting with the low entropy and robust reasoning of ChatGPT models. The study further demonstrates that model size is not the sole determinant of performance, as evidenced by Grok 3 underperformance despite its substantial parameter count. By employing multi-image contexts, rejection mechanisms, and entropy-based consistency metrics, this benchmark sets a new standard for evaluating multimodal LLMs, enabling a more robust and reliable assessment of next-generation AI systems.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2502.16732v2",
"version_id": "http://arxiv.org/abs/2502.16732v2",
"title": "DeepSeek reshaping healthcare in China's tertiary hospitals",
"published": "2025-02-23T22:09:17Z",
"updated": "2025-02-27T10:24:58Z",
"authors": [
"Jishizhan Chen",
"Qingzeng Zhang"
],
"summary": "The rapid integration of artificial intelligence (AI) into healthcare is transforming clinical decision-making and hospital operations. DeepSeek has emerged as a leading AI system, widely deployed across China's tertiary hospitals since January 2025. Initially implemented in Shanghai's major medical institutions, it has since expanded nationwide, enhancing diagnostic accuracy, streamlining workflows, and improving patient management. AI-powered pathology, imaging analysis, and clinical decision support systems have demonstrated significant potential in optimizing medical processes and reducing the cognitive burden on healthcare professionals. However, the widespread adoption of AI in healthcare raises critical regulatory and ethical challenges, particularly regarding accountability in AI-assisted diagnosis and the risk of automation bias. The absence of a well-defined liability framework underscores the need for policies that ensure AI functions as an assistive tool rather than an autonomous decision-maker. With continued technological advancements, AI is expected to integrate multimodal data sources, such as genomics and radiomics, paving the way for precision medicine and personalized treatment strategies. The future of AI in healthcare depends on the development of transparent regulatory structures, industry collaboration, and adaptive governance frameworks that balance innovation with responsibility, ensuring equitable and effective AI-driven medical services.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CY"
},
{
"id": "2502.17764v2",
"version_id": "http://arxiv.org/abs/2502.17764v2",
"title": "DeepSeek vs. ChatGPT vs. Claude: A Comparative Study for Scientific Computing and Scientific Machine Learning Tasks",
"published": "2025-02-25T01:49:50Z",
"updated": "2025-03-08T22:55:31Z",
"authors": [
"Qile Jiang",
"Zhiwei Gao",
"George Em Karniadakis"
],
"summary": "Large Language Models (LLMs) have emerged as powerful tools for tackling a wide range of problems, including those in scientific computing, particularly in solving partial differential equations (PDEs). However, different models exhibit distinct strengths and preferences, resulting in varying levels of performance. In this paper, we compare the capabilities of the most advanced LLMs--DeepSeek, ChatGPT, and Claude--along with their reasoning-optimized versions in addressing computational challenges. Specifically, we evaluate their proficiency in solving traditional numerical problems in scientific computing as well as leveraging scientific machine learning techniques for PDE-based problems. We designed all our experiments so that a non-trivial decision is required, e.g. defining the proper space of input functions for neural operator learning. Our findings show that reasoning and hybrid-reasoning models consistently and significantly outperform non-reasoning ones in solving challenging problems, with ChatGPT o3-mini-high generally offering the fastest reasoning speed.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.LG"
},
{
"id": "2502.17947v1",
"version_id": "http://arxiv.org/abs/2502.17947v1",
"title": "DeepSeek-R1 Outperforms Gemini 2.0 Pro, OpenAI o1, and o3-mini in Bilingual Complex Ophthalmology Reasoning",
"published": "2025-02-25T08:08:53Z",
"updated": "2025-02-25T08:08:53Z",
"authors": [
"Pusheng Xu",
"Yue Wu",
"Kai Jin",
"Xiaolan Chen",
"Mingguang He",
"Danli Shi"
],
"summary": "Purpose: To evaluate the accuracy and reasoning ability of DeepSeek-R1 and three other recently released large language models (LLMs) in bilingual complex ophthalmology cases. Methods: A total of 130 multiple-choice questions (MCQs) related to diagnosis (n = 39) and management (n = 91) were collected from the Chinese ophthalmology senior professional title examination and categorized into six topics. These MCQs were translated into English using DeepSeek-R1. The responses of DeepSeek-R1, Gemini 2.0 Pro, OpenAI o1 and o3-mini were generated under default configurations between February 15 and February 20, 2025. Accuracy was calculated as the proportion of correctly answered questions, with omissions and extra answers considered incorrect. Reasoning ability was evaluated through analyzing reasoning logic and the causes of reasoning error. Results: DeepSeek-R1 demonstrated the highest overall accuracy, achieving 0.862 in Chinese MCQs and 0.808 in English MCQs. Gemini 2.0 Pro, OpenAI o1, and OpenAI o3-mini attained accuracies of 0.715, 0.685, and 0.692 in Chinese MCQs (all P<0.001 compared with DeepSeek-R1), and 0.746 (P=0.115), 0.723 (P=0.027), and 0.577 (P<0.001) in English MCQs, respectively. DeepSeek-R1 achieved the highest accuracy across five topics in both Chinese and English MCQs. It also excelled in management questions conducted in Chinese (all P<0.05). Reasoning ability analysis showed that the four LLMs shared similar reasoning logic. Ignoring key positive history, ignoring key positive signs, misinterpretation medical data, and too aggressive were the most common causes of reasoning errors. Conclusion: DeepSeek-R1 demonstrated superior performance in bilingual complex ophthalmology reasoning tasks than three other state-of-the-art LLMs. While its clinical applicability remains challenging, it shows promise for supporting diagnosis and clinical decision-making.",
"comment": "29 pages, 4 figures, 1 table",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2502.18210v1",
"version_id": "http://arxiv.org/abs/2502.18210v1",
"title": "From ChatGPT to DeepSeek: Can LLMs Simulate Humanity?",
"published": "2025-02-25T13:54:47Z",
"updated": "2025-02-25T13:54:47Z",
"authors": [
"Qian Wang",
"Zhenheng Tang",
"Bingsheng He"
],
"summary": "Simulation powered by Large Language Models (LLMs) has become a promising method for exploring complex human social behaviors. However, the application of LLMs in simulations presents significant challenges, particularly regarding their capacity to accurately replicate the complexities of human behaviors and societal dynamics, as evidenced by recent studies highlighting discrepancies between simulated and real-world interactions. We rethink LLM-based simulations by emphasizing both their limitations and the necessities for advancing LLM simulations. By critically examining these challenges, we aim to offer actionable insights and strategies for enhancing the applicability of LLM simulations in human society in the future.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CY"
},
{
"id": "2503.04783v1",
"version_id": "http://arxiv.org/abs/2503.04783v1",
"title": "Comparative Analysis Based on DeepSeek, ChatGPT, and Google Gemini: Features, Techniques, Performance, Future Prospects",
"published": "2025-02-25T19:55:35Z",
"updated": "2025-02-25T19:55:35Z",
"authors": [
"Anichur Rahman",
"Shahariar Hossain Mahir",
"Md Tanjum An Tashrif",
"Airin Afroj Aishi",
"Md Ahsan Karim",
"Dipanjali Kundu",
"Tanoy Debnath",
"Md. Abul Ala Moududi",
"MD. Zunead Abedin Eidmum"
],
"summary": "Nowadays, DeepSeek, ChatGPT, and Google Gemini are the most trending and exciting Large Language Model (LLM) technologies for reasoning, multimodal capabilities, and general linguistic performance worldwide. DeepSeek employs a Mixture-of-Experts (MoE) approach, activating only the parameters most relevant to the task at hand, which makes it especially effective for domain-specific work. On the other hand, ChatGPT relies on a dense transformer model enhanced through reinforcement learning from human feedback (RLHF), and then Google Gemini actually uses a multimodal transformer architecture that integrates text, code, and images into a single framework. However, by using those technologies, people can be able to mine their desired text, code, images, etc, in a cost-effective and domain-specific inference. People may choose those techniques based on the best performance. In this regard, we offer a comparative study based on the DeepSeek, ChatGPT, and Gemini techniques in this research. Initially, we focus on their methods and materials, appropriately including the data selection criteria. Then, we present state-of-the-art features of DeepSeek, ChatGPT, and Gemini based on their applications. Most importantly, we show the technological comparison among them and also cover the dataset analysis for various applications. Finally, we address extensive research areas and future potential guidance regarding LLM-based AI research for the community.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2503.00624v1",
"version_id": "http://arxiv.org/abs/2503.00624v1",
"title": "An evaluation of DeepSeek Models in Biomedical Natural Language Processing",
"published": "2025-03-01T21:26:29Z",
"updated": "2025-03-01T21:26:29Z",
"authors": [
"Zaifu Zhan",
"Shuang Zhou",
"Huixue Zhou",
"Jiawen Deng",
"Yu Hou",
"Jeremy Yeung",
"Rui Zhang"
],
"summary": "The advancement of Large Language Models (LLMs) has significantly impacted biomedical Natural Language Processing (NLP), enhancing tasks such as named entity recognition, relation extraction, event extraction, and text classification. In this context, the DeepSeek series of models have shown promising potential in general NLP tasks, yet their capabilities in the biomedical domain remain underexplored. This study evaluates multiple DeepSeek models (Distilled-DeepSeek-R1 series and Deepseek-LLMs) across four key biomedical NLP tasks using 12 datasets, benchmarking them against state-of-the-art alternatives (Llama3-8B, Qwen2.5-7B, Mistral-7B, Phi-4-14B, Gemma-2-9B). Our results reveal that while DeepSeek models perform competitively in named entity recognition and text classification, challenges persist in event and relation extraction due to precision-recall trade-offs. We provide task-specific model recommendations and highlight future research directions. This evaluation underscores the strengths and limitations of DeepSeek models in biomedical NLP, guiding their future deployment and optimization.",
"comment": "Plan to submit to AMIA 2025 Annual Symposium. 10 pages",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2503.01923v1",
"version_id": "http://arxiv.org/abs/2503.01923v1",
"title": "Output Length Effect on DeepSeek-R1's Safety in Forced Thinking",
"published": "2025-03-02T06:29:22Z",
"updated": "2025-03-02T06:29:22Z",
"authors": [
"Xuying Li",
"Zhuo Li",
"Yuji Kosuga",
"Victor Bian"
],
"summary": "Large Language Models (LLMs) have demonstrated strong reasoning capabilities, but their safety under adversarial conditions remains a challenge. This study examines the impact of output length on the robustness of DeepSeek-R1, particularly in Forced Thinking scenarios. We analyze responses across various adversarial prompts and find that while longer outputs can improve safety through self-correction, certain attack types exploit extended generations. Our findings suggest that output length should be dynamically controlled to balance reasoning effectiveness and security. We propose reinforcement learning-based policy adjustments and adaptive token length regulation to enhance LLM safety.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2503.02032v1",
"version_id": "http://arxiv.org/abs/2503.02032v1",
"title": "Comparative Analysis of OpenAI GPT-4o and DeepSeek R1 for Scientific Text Categorization Using Prompt Engineering",
"published": "2025-03-03T20:09:35Z",
"updated": "2025-03-03T20:09:35Z",
"authors": [
"Aniruddha Maiti",
"Samuel Adewumi",
"Temesgen Alemayehu Tikure",
"Zichun Wang",
"Niladri Sengupta",
"Anastasiia Sukhanova",
"Ananya Jana"
],
"summary": "This study examines how large language models categorize sentences from scientific papers using prompt engineering. We use two advanced web-based models, GPT-4o (by OpenAI) and DeepSeek R1, to classify sentences into predefined relationship categories. DeepSeek R1 has been tested on benchmark datasets in its technical report. However, its performance in scientific text categorization remains unexplored. To address this gap, we introduce a new evaluation method designed specifically for this task. We also compile a dataset of cleaned scientific papers from diverse domains. This dataset provides a platform for comparing the two models. Using this dataset, we analyze their effectiveness and consistency in categorization.",
"comment": "Accepted to ASEE North Central Section 2025",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2503.09956v4",
"version_id": "http://arxiv.org/abs/2503.09956v4",
"title": "DeepSeek-Inspired Exploration of RL-based LLMs and Synergy with Wireless Networks: A Survey",
"published": "2025-03-13T01:59:11Z",
"updated": "2025-10-20T06:44:00Z",
"authors": [
"Yu Qiao",
"Phuong-Nam Tran",
"Ji Su Yoon",
"Loc X. Nguyen",
"Eui-Nam Huh",
"Dusit Niyato",
"Choong Seon Hong"
],
"summary": "Reinforcement learning (RL)-based large language models (LLMs), such as ChatGPT, DeepSeek, and Grok-3, have attracted widespread attention for their remarkable capabilities in multimodal data understanding. Meanwhile, the rapid expansion of information services has led to a growing demand for AI-enabled wireless networks. The open-source DeepSeek models are famous for their innovative designs, such as large-scale pure RL and cost-efficient training, which make them well-suited for practical deployment in wireless networks. By integrating DeepSeek-style LLMs with wireless infrastructures, a synergistic opportunity arises: the DeepSeek-style LLMs enhance network optimization with strong reasoning and decision-making abilities, while wireless infrastructure enables the broad deployment of these models. Motivated by this convergence, this survey presents a comprehensive DeepSeek-inspired exploration of RL-based LLMs in the context of wireless networks. We begin by reviewing key techniques behind network optimization to establish a foundation for understanding DeepSeek-style LLM integration. Next, we examine recent advancements in RL-based LLMs, using DeepSeek models as a representative example. Building on this, we explore the synergy between the two domains, highlighting motivations, challenges, and potential solutions. Finally, we highlight emerging directions for integrating LLMs with wireless networks, such as quantum, on-device, and neural-symbolic LLM models, as well as embodied AI agents. Overall, this survey offers a comprehensive examination of the interplay between DeepSeek-style LLMs and wireless networks, demonstrating how these domains can mutually enhance each other to drive innovation.",
"comment": "45 pages, 12 figures",
"journal_ref": "ACM Computing Surveys, Nov. 2025",
"doi": "10.1145/3776745",
"primary": "cs.LG"
},
{
"id": "2503.10486v2",
"version_id": "http://arxiv.org/abs/2503.10486v2",
"title": "LLMs in Disease Diagnosis: A Comparative Study of DeepSeek-R1 and O3 Mini Across Chronic Health Conditions",
"published": "2025-03-13T15:54:26Z",
"updated": "2025-06-20T02:33:03Z",
"authors": [
"Gaurav Kumar Gupta",
"Pranal Pande",
"Nirajan Acharya",
"Aniket Kumar Singh",
"Suman Niroula"
],
"summary": "Large Language Models (LLMs) are revolutionizing medical diagnostics by enhancing both disease classification and clinical decision-making. In this study, we evaluate the performance of two LLM- based diagnostic tools, DeepSeek R1 and O3 Mini, using a structured dataset of symptoms and diagnoses. We assessed their predictive accuracy at both the disease and category levels, as well as the reliability of their confidence scores. DeepSeek R1 achieved a disease-level accuracy of 76% and an overall accuracy of 82%, outperforming O3 Mini, which attained 72% and 75% respectively. Notably, DeepSeek R1 demonstrated exceptional performance in Mental Health, Neurological Disorders, and Oncology, where it reached 100% accuracy, while O3 Mini excelled in Autoimmune Disease classification with 100% accuracy. Both models, however, struggled with Respiratory Disease classification, recording accuracies of only 40% for DeepSeek R1 and 20% for O3 Mini. Additionally, the analysis of confidence scores revealed that DeepSeek R1 provided high-confidence predictions in 92% of cases, compared to 68% for O3 Mini. Ethical considerations regarding bias, model interpretability, and data privacy are also discussed to ensure the responsible integration of LLMs into clinical practice. Overall, our findings offer valuable insights into the strengths and limitations of LLM-based diagnostic systems and provide a roadmap for future enhancements in AI-driven healthcare.",
"comment": "12 pages, 3 figures",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2503.11068v2",
"version_id": "http://arxiv.org/abs/2503.11068v2",
"title": "DeepSeek Powered Solid Dosage Formulation Design and Development",
"published": "2025-03-14T04:23:59Z",
"updated": "2025-03-21T08:25:11Z",
"authors": [
"Leqi Lin",
"Xingyu Zhou",
"Kaiyuan Yang",
"Xizhong Chen"
],
"summary": "Pharmaceutical process design and development for generic, innovative, or personalized drugs have always been a time-consuming, costly, rigorous process, that involves multi-stage evaluation for better quality control and assurance. Large language models (LLMs), a type of generative artificial intelligence system, can augment laboratory research in the pharmaceutical engineering process by helping scientists to extract knowledge from literature, design parameters, and collect and interpret experimental data ultimately accelerating scientific discovery. LLMs with prompt engineering technologies change the researchers thinking protocol from traditional empirical knowledge to streamlined thinking that connects the performance and structured parameters together. In this work, we investigate and evaluate how prompt engineering technologies can enhance the drug design process from different strategies such as zero-shot, few-shot, chain-of-thought, etc. The dissolution profile for specific drugs is predicted and suggested from the LLMs model. Furthermore, the fundamental physical properties such as PSD, aspect ratio, and specific surface area could be inversely designed from the LLMs model. Finally, all the results are evaluated and validated by real-world cases to prove the reliability of prompt engineering techniques. Initial evaluations show an MSE of 23.61 and R2 of 0.97 in zero-shot, an MSE of 114.89 and R2 of 0.90 in zero-shot-CoT, an MSE of 57.0 and R2 of 0.92 in few-shot, a MSE of 22.56 and R2 of 0.97 in few-shot-CoT and a MSE of 10.56 and R2 of 0.99 with the involvement of RAG. This work breaks down any barriers in developing a systematic framework where LLMs assist in formulation design, process control, and decision-making. Finally, we conclude the work by discussing open challenges and future research directions in pharmaceutical processes.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.ET"
},
{
"id": "2503.11486v1",
"version_id": "http://arxiv.org/abs/2503.11486v1",
"title": "A Review of DeepSeek Models' Key Innovative Techniques",
"published": "2025-03-14T15:11:29Z",
"updated": "2025-03-14T15:11:29Z",
"authors": [
"Chengen Wang",
"Murat Kantarcioglu"
],
"summary": "DeepSeek-V3 and DeepSeek-R1 are leading open-source Large Language Models (LLMs) for general-purpose tasks and reasoning, achieving performance comparable to state-of-the-art closed-source models from companies like OpenAI and Anthropic -- while requiring only a fraction of their training costs. Understanding the key innovative techniques behind DeepSeek's success is crucial for advancing LLM research. In this paper, we review the core techniques driving the remarkable effectiveness and efficiency of these models, including refinements to the transformer architecture, innovations such as Multi-Head Latent Attention and Mixture of Experts, Multi-Token Prediction, the co-design of algorithms, frameworks, and hardware, the Group Relative Policy Optimization algorithm, post-training with pure reinforcement learning and iterative training alternating between supervised fine-tuning and reinforcement learning. Additionally, we identify several open questions and highlight potential research opportunities in this rapidly advancing field.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.LG"
},
{
"id": "2504.03665v2",
"version_id": "http://arxiv.org/abs/2504.03665v2",
"title": "LLM & HPC:Benchmarking DeepSeek's Performance in High-Performance Computing Tasks",
"published": "2025-03-15T21:42:54Z",
"updated": "2025-04-28T20:31:31Z",
"authors": [
"Noujoud Nader",
"Patrick Diehl",
"Steve Brandt",
"Hartmut Kaiser"
],
"summary": "Large Language Models (LLMs), such as GPT-4 and DeepSeek, have been applied to a wide range of domains in software engineering. However, their potential in the context of High-Performance Computing (HPC) much remains to be explored. This paper evaluates how well DeepSeek, a recent LLM, performs in generating a set of HPC benchmark codes: a conjugate gradient solver, the parallel heat equation, parallel matrix multiplication, DGEMM, and the STREAM triad operation. We analyze DeepSeek's code generation capabilities for traditional HPC languages like Cpp, Fortran, Julia and Python. The evaluation includes testing for code correctness, performance, and scaling across different configurations and matrix sizes. We also provide a detailed comparison between DeepSeek and another widely used tool: GPT-4. Our results demonstrate that while DeepSeek generates functional code for HPC tasks, it lags behind GPT-4, in terms of scalability and execution efficiency of the generated code.",
"comment": "9 pages, 2 figures, 3 tables, conference",
"journal_ref": "",
"doi": "10.1007/978-3-032-07612-0_48",
"primary": "cs.DC"
},
{
"id": "2503.13549v1",
"version_id": "http://arxiv.org/abs/2503.13549v1",
"title": "A Showdown of ChatGPT vs DeepSeek in Solving Programming Tasks",
"published": "2025-03-16T14:35:36Z",
"updated": "2025-03-16T14:35:36Z",
"authors": [
"Ronas Shakya",
"Sam Urmian",
"Mohammad Khalil"
],
"summary": "The advancement of large language models (LLMs) has created a competitive landscape for AI-assisted programming tools. This study evaluates two leading models: ChatGPT 03-mini and DeepSeek-R1 on their ability to solve competitive programming tasks from Codeforces. Using 29 programming tasks of three levels of easy, medium, and hard difficulty, we assessed the outcome of both models by their accepted solutions, memory efficiency, and runtime performance. Our results indicate that while both models perform similarly on easy tasks, ChatGPT outperforms DeepSeek-R1 on medium-difficulty tasks, achieving a 54.5% success rate compared to DeepSeek 18.1%. Both models struggled with hard tasks, thus highlighting some ongoing challenges LLMs face in handling highly complex programming problems. These findings highlight key differences in both model capabilities and their computational power, offering valuable insights for developers and researchers working to advance AI-driven programming tools.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.SE"
},
{
"id": "2503.16529v2",
"version_id": "http://arxiv.org/abs/2503.16529v2",
"title": "Safety Evaluation and Enhancement of DeepSeek Models in Chinese Contexts",
"published": "2025-03-18T08:38:10Z",
"updated": "2025-05-16T13:29:19Z",
"authors": [
"Wenjing Zhang",
"Xuejiao Lei",
"Zhaoxiang Liu",
"Limin Han",
"Jiaojiao Zhao",
"Junting Guo",
"Zhenhong Long",
"Shu Yang",
"Meijuan An",
"Beibei Huang",
"Rongjia Du",
"Ning Wang",
"Kai Wang",
"Shiguo Lian"
],
"summary": "DeepSeek-R1, renowned for its exceptional reasoning capabilities and open-source strategy, is significantly influencing the global artificial intelligence landscape. However, it exhibits notable safety shortcomings. Recent research conducted by Robust Intelligence, a subsidiary of Cisco, in collaboration with the University of Pennsylvania, revealed that DeepSeek-R1 achieves a 100\\% attack success rate when processing harmful prompts. Furthermore, multiple security firms and research institutions have identified critical security vulnerabilities within the model. Although China Unicom has uncovered safety vulnerabilities of R1 in Chinese contexts, the safety capabilities of the remaining distilled models in the R1 series have not yet been comprehensively evaluated. To address this gap, this study utilizes the comprehensive Chinese safety benchmark CHiSafetyBench to conduct an in-depth safety evaluation of the DeepSeek-R1 series distilled models. The objective is to assess the safety capabilities of these models in Chinese contexts both before and after distillation, and to further elucidate the adverse effects of distillation on model safety. Building on these findings, we implement targeted safety enhancements for the entire DeepSeek-R1 model series. Evaluation results indicate that the enhanced models achieve significant improvements in safety while maintaining reasoning capabilities without notable degradation. We open-source the safety-enhanced models at https://github.com/UnicomAI/DeepSeek-R1-Safe to serve as a valuable resource for future research and optimization of DeepSeek models.",
"comment": "21 pages, 13 figures, 4 tables",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2503.15092v1",
"version_id": "http://arxiv.org/abs/2503.15092v1",
"title": "Towards Understanding the Safety Boundaries of DeepSeek Models: Evaluation and Findings",
"published": "2025-03-19T10:44:37Z",
"updated": "2025-03-19T10:44:37Z",
"authors": [
"Zonghao Ying",
"Guangyi Zheng",
"Yongxin Huang",
"Deyue Zhang",
"Wenxin Zhang",
"Quanchen Zou",
"Aishan Liu",
"Xianglong Liu",
"Dacheng Tao"
],
"summary": "This study presents the first comprehensive safety evaluation of the DeepSeek models, focusing on evaluating the safety risks associated with their generated content. Our evaluation encompasses DeepSeek's latest generation of large language models, multimodal large language models, and text-to-image models, systematically examining their performance regarding unsafe content generation. Notably, we developed a bilingual (Chinese-English) safety evaluation dataset tailored to Chinese sociocultural contexts, enabling a more thorough evaluation of the safety capabilities of Chinese-developed models. Experimental results indicate that despite their strong general capabilities, DeepSeek models exhibit significant safety vulnerabilities across multiple risk dimensions, including algorithmic discrimination and sexual content. These findings provide crucial insights for understanding and improving the safety of large foundation models. Our code is available at https://github.com/NY1024/DeepSeek-Safety-Eval.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CR"
},
{
"id": "2503.16040v2",
"version_id": "http://arxiv.org/abs/2503.16040v2",
"title": "Evaluating Test-Time Scaling LLMs for Legal Reasoning: OpenAI o1, DeepSeek-R1, and Beyond",
"published": "2025-03-20T11:14:39Z",
"updated": "2025-11-10T11:28:29Z",
"authors": [
"Yinghao Hu",
"Yaoyao Yu",
"Leilei Gan",
"Bin Wei",
"Kun Kuang",
"Fei Wu"
],
"summary": "Recent advances in test-time scaling of large language models (LLMs), exemplified by DeepSeek-R1 and OpenAI's o1, show that extending the chain of thought during inference can significantly improve general reasoning performance. However, the impact of this paradigm on legal reasoning remains insufficiently explored. To address this gap, we present the first systematic evaluation of 12 LLMs, including both reasoning-focused and general-purpose models, across 17 Chinese and English legal tasks spanning statutory and case-law traditions. In addition, we curate a bilingual chain-of-thought dataset for legal reasoning through distillation from DeepSeek-R1 and develop Legal-R1, an open-source model specialized for the legal domain. Experimental results show that Legal-R1 delivers competitive performance across diverse tasks. DeepSeek-R1 exhibits clear advantages in Chinese legal reasoning, while OpenAI's o1 achieves comparable results on English tasks. We further conduct a detailed error analysis, which reveals recurring issues such as outdated legal knowledge, limited capacity for legal interpretation, and susceptibility to factual hallucinations. These findings delineate the main obstacles confronting legal-domain LLMs and suggest promising directions for future research.",
"comment": "23 pages, Published in Findings of the Association for Computational Linguistics: EMNLP 2025",
"journal_ref": "",
"doi": "10.18653/v1/2025.findings-emnlp.742",
"primary": "cs.CL"
},
{
"id": "2503.16304v3",
"version_id": "http://arxiv.org/abs/2503.16304v3",
"title": "Bridging Technology and Humanities: Evaluating the Impact of Large Language Models on Social Sciences Research with DeepSeek-R1",
"published": "2025-03-20T16:25:24Z",
"updated": "2025-04-15T15:09:24Z",
"authors": [
"Peiran Gu",
"Fuhao Duan",
"Wenhao Li",
"Bochen Xu",
"Ying Cai",
"Teng Yao",
"Chenxun Zhuo",
"Tianming Liu",
"Bao Ge"
],
"summary": "In recent years, the development of Large Language Models (LLMs) has made significant breakthroughs in the field of natural language processing and has gradually been applied to the field of humanities and social sciences research. LLMs have a wide range of application value in the field of humanities and social sciences because of its strong text understanding, generation and reasoning capabilities. In humanities and social sciences research, LLMs can analyze large-scale text data and make inferences. This article analyzes the large language model DeepSeek-R1 from seven aspects: low-resource language translation, educational question-answering, student writing improvement in higher education, logical reasoning, educational measurement and psychometrics, public health policy analysis, and art education . Then we compare the answers given by DeepSeek-R1 in the seven aspects with the answers given by o1-preview. DeepSeek-R1 performs well in the humanities and social sciences, answering most questions correctly and logically, and can give reasonable analysis processes and explanations. Compared with o1-preview, it can automatically generate reasoning processes and provide more detailed explanations, which is suitable for beginners or people who need to have a detailed understanding of this knowledge, while o1-preview is more suitable for quick reading. Through analysis, it is found that LLM has broad application potential in the field of humanities and social sciences, and shows great advantages in improving text analysis efficiency, language communication and other fields. LLM's powerful language understanding and generation capabilities enable it to deeply explore complex problems in the field of humanities and social sciences, and provide innovative tools for academic research and practical applications.",
"comment": "52 pages, 19 figures",
"journal_ref": "",
"doi": "",
"primary": "cs.CY"
},
{
"id": "2504.00016v1",
"version_id": "http://arxiv.org/abs/2504.00016v1",
"title": "Medical Reasoning in LLMs: An In-Depth Analysis of DeepSeek R1",
"published": "2025-03-27T09:18:08Z",
"updated": "2025-03-27T09:18:08Z",
"authors": [
"Birger Moell",
"Fredrik Sand Aronsson",
"Sanian Akbar"
],
"summary": "Integrating large language models (LLMs) like DeepSeek R1 into healthcare requires rigorous evaluation of their reasoning alignment with clinical expertise. This study assesses DeepSeek R1's medical reasoning against expert patterns using 100 MedQA clinical cases. The model achieved 93% diagnostic accuracy, demonstrating systematic clinical judgment through differential diagnosis, guideline-based treatment selection, and integration of patient-specific factors. However, error analysis of seven incorrect cases revealed persistent limitations: anchoring bias, challenges reconciling conflicting data, insufficient exploration of alternatives, overthinking, knowledge gaps, and premature prioritization of definitive treatment over intermediate care. Crucially, reasoning length correlated with accuracy - shorter responses (<5,000 characters) were more reliable, suggesting extended explanations may signal uncertainty or rationalization of errors. While DeepSeek R1 exhibits foundational clinical reasoning capabilities, recurring flaws highlight critical areas for refinement, including bias mitigation, knowledge updates, and structured reasoning frameworks. These findings underscore LLMs' potential to augment medical decision-making through artificial reasoning but emphasize the need for domain-specific validation, interpretability safeguards, and confidence metrics (e.g., response length thresholds) to ensure reliability in real-world applications.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2503.23130v3",
"version_id": "http://arxiv.org/abs/2503.23130v3",
"title": "Can DeepSeek Reason Like a Surgeon? An Empirical Evaluation for Vision-Language Understanding in Robotic-Assisted Surgery",
"published": "2025-03-29T15:48:46Z",
"updated": "2025-04-04T02:45:12Z",
"authors": [
"Boyi Ma",
"Yanguang Zhao",
"Jie Wang",
"Guankun Wang",
"Kun Yuan",
"Tong Chen",
"Long Bai",
"Hongliang Ren"
],
"summary": "The DeepSeek models have shown exceptional performance in general scene understanding, question-answering (QA), and text generation tasks, owing to their efficient training paradigm and strong reasoning capabilities. In this study, we investigate the dialogue capabilities of the DeepSeek model in robotic surgery scenarios, focusing on tasks such as Single Phrase QA, Visual QA, and Detailed Description. The Single Phrase QA tasks further include sub-tasks such as surgical instrument recognition, action understanding, and spatial position analysis. We conduct extensive evaluations using publicly available datasets, including EndoVis18 and CholecT50, along with their corresponding dialogue data. Our empirical study shows that, compared to existing general-purpose multimodal large language models, DeepSeek-VL2 performs better on complex understanding tasks in surgical scenes. Additionally, although DeepSeek-V3 is purely a language model, we find that when image tokens are directly inputted, the model demonstrates better performance on single-sentence QA tasks. However, overall, the DeepSeek models still fall short of meeting the clinical requirements for understanding surgical scenes. Under general prompts, DeepSeek models lack the ability to effectively analyze global surgical concepts and fail to provide detailed insights into surgical scenarios. Based on our observations, we argue that the DeepSeek models are not ready for vision-language tasks in surgical contexts without fine-tuning on surgery-specific datasets.",
"comment": "Technical Report",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2504.00421v1",
"version_id": "http://arxiv.org/abs/2504.00421v1",
"title": "Can LLMs Assist Computer Education? an Empirical Case Study of DeepSeek",
"published": "2025-04-01T04:58:16Z",
"updated": "2025-04-01T04:58:16Z",
"authors": [
"Dongfu Xiao",
"Chen Gao",
"Zhengquan Luo",
"Chi Liu",
"Sheng Shen"
],
"summary": "This study presents an empirical case study to assess the efficacy and reliability of DeepSeek-V3, an emerging large language model, within the context of computer education. The evaluation employs both CCNA simulation questions and real-world inquiries concerning computer network security posed by Chinese network engineers. To ensure a thorough evaluation, diverse dimensions are considered, encompassing role dependency, cross-linguistic proficiency, and answer reproducibility, accompanied by statistical analysis. The findings demonstrate that the model performs consistently, regardless of whether prompts include a role definition or not. In addition, its adaptability across languages is confirmed by maintaining stable accuracy in both original and translated datasets. A distinct contrast emerges between its performance on lower-order factual recall tasks and higher-order reasoning exercises, which underscores its strengths in retrieving information and its limitations in complex analytical tasks. Although DeepSeek-V3 offers considerable practical value for network security education, challenges remain in its capability to process multimodal data and address highly intricate topics. These results provide valuable insights for future refinement of large language models in specialized professional environments.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2504.07128v3",
"version_id": "http://arxiv.org/abs/2504.07128v3",
"title": "DeepSeek-R1 Thoughtology: Let's think about LLM Reasoning",
"published": "2025-04-02T00:36:08Z",
"updated": "2026-01-15T21:08:10Z",
"authors": [
"Sara Vera Marjanović",
"Arkil Patel",
"Vaibhav Adlakha",
"Milad Aghajohari",
"Parishad BehnamGhader",
"Mehar Bhatia",
"Aditi Khandelwal",
"Austin Kraft",
"Benno Krojer",
"Xing Han Lù",
"Nicholas Meade",
"Dongchan Shin",
"Amirhossein Kazemnejad",
"Gaurav Kamath",
"Marius Mosbach",
"Karolina Stańczak",
"Siva Reddy"
],
"summary": "Large Reasoning Models like DeepSeek-R1 mark a fundamental shift in how LLMs approach complex problems. Instead of directly producing an answer for a given input, DeepSeek-R1 creates detailed multi-step reasoning chains, seemingly \"thinking\" about a problem before providing an answer. This reasoning process is publicly available to the user, creating endless opportunities for studying the reasoning behaviour of the model and opening up the field of Thoughtology. Starting from a taxonomy of DeepSeek-R1's basic building blocks of reasoning, our analyses on DeepSeek-R1 investigate the impact and controllability of thought length, management of long or confusing contexts, cultural and safety concerns, and the status of DeepSeek-R1 vis-à-vis cognitive phenomena, such as human-like language processing and world modelling. Our findings paint a nuanced picture. Notably, we show DeepSeek-R1 has a 'sweet spot' of reasoning, where extra inference time can impair model performance. Furthermore, we find a tendency for DeepSeek-R1 to persistently ruminate on previously explored problem formulations, obstructing further exploration. We also note strong safety vulnerabilities of DeepSeek-R1 compared to its non-reasoning counterpart, which can also compromise safety-aligned LLMs.",
"comment": "135 pages, Published to TMLR",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2504.02888v1",
"version_id": "http://arxiv.org/abs/2504.02888v1",
"title": "A Status Quo Investigation of Large Language Models towards Cost-Effective CFD Automation with OpenFOAMGPT: ChatGPT vs. Qwen vs. Deepseek",
"published": "2025-04-02T14:04:52Z",
"updated": "2025-04-02T14:04:52Z",
"authors": [
"Wenkang Wang",
"Ran Xu",
"Jingsen Feng",
"Qingfu Zhang",
"Xu Chu"
],
"summary": "We evaluated the performance of OpenFOAMGPT incorporating multiple large-language models. Some of the present models efficiently manage different CFD tasks such as adjusting boundary conditions, turbulence models, and solver configurations, although their token cost and stability vary. Locally deployed smaller models like QwQ-32B struggled with generating valid solver files for complex processes. Zero-shot prompting commonly failed in simulations with intricate settings, even for large models. Challenges with boundary conditions and solver keywords stress the requirement for expert supervision, indicating that further development is needed to fully automate specialized CFD simulations.",
"comment": "",
"journal_ref": "",
"doi": "10.1016/j.taml.2025.100623",
"primary": "cs.CL"
},
{
"id": "2504.03219v1",
"version_id": "http://arxiv.org/abs/2504.03219v1",
"title": "From ChatGPT to DeepSeek AI: A Comprehensive Analysis of Evolution, Deviation, and Future Implications in AI-Language Models",
"published": "2025-04-04T07:08:29Z",
"updated": "2025-04-04T07:08:29Z",
"authors": [
"Simrandeep Singh",
"Shreya Bansal",
"Abdulmotaleb El Saddik",
"Mukesh Saini"
],
"summary": "The rapid advancement of artificial intelligence (AI) has reshaped the field of natural language processing (NLP), with models like OpenAI ChatGPT and DeepSeek AI. Although ChatGPT established a strong foundation for conversational AI, DeepSeek AI introduces significant improvements in architecture, performance, and ethical considerations. This paper presents a detailed analysis of the evolution from ChatGPT to DeepSeek AI, highlighting their technical differences, practical applications, and broader implications for AI development. To assess their capabilities, we conducted a case study using a predefined set of multiple choice questions in various domains, evaluating the strengths and limitations of each model. By examining these aspects, we provide valuable insight into the future trajectory of AI, its potential to transform industries, and key research directions for improving AI-driven language models.",
"comment": "10 pages, 1 figure, 4 tables",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2504.08120v3",
"version_id": "http://arxiv.org/abs/2504.08120v3",
"title": "DeepSeek-R1 vs. o3-mini: How Well can Reasoning LLMs Evaluate MT and Summarization?",
"published": "2025-04-10T20:39:18Z",
"updated": "2025-05-30T12:59:36Z",
"authors": [
"Daniil Larionov",
"Sotaro Takeshita",
"Ran Zhang",
"Yanran Chen",
"Christoph Leiter",
"Zhipin Wang",
"Christian Greisinger",
"Steffen Eger"
],
"summary": "Reasoning-enabled large language models (LLMs) excel in logical tasks, yet their utility for evaluating natural language generation remains unexplored. This study systematically compares reasoning LLMs with non-reasoning counterparts across machine translation and text summarization evaluation tasks. We evaluate eight models spanning state-of-the-art reasoning models (DeepSeek-R1, OpenAI o3), their distilled variants (8B-70B parameters), and equivalent non-reasoning LLMs. Experiments on WMT23 and SummEval benchmarks reveal architecture and task-dependent benefits: OpenAI o3-mini models show improved performance with increased reasoning on MT, while DeepSeek-R1 and generally underperforms compared to its non-reasoning variant except in summarization consistency evaluation. Correlation analysis demonstrates that reasoning token usage correlates with evaluation quality only in specific models, while almost all models generally allocate more reasoning tokens when identifying more quality issues. Distillation maintains reasonable performance up to 32B parameter models but degrades substantially at 8B scale. This work provides the first assessment of reasoning LLMs for NLG evaluation and comparison to non-reasoning models. We share our code to facilitate further research: https://github.com/NL2G/reasoning-eval.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2504.10081v1",
"version_id": "http://arxiv.org/abs/2504.10081v1",
"title": "RealSafe-R1: Safety-Aligned DeepSeek-R1 without Compromising Reasoning Capability",
"published": "2025-04-14T10:26:37Z",
"updated": "2025-04-14T10:26:37Z",
"authors": [
"Yichi Zhang",
"Zihao Zeng",
"Dongbai Li",
"Yao Huang",
"Zhijie Deng",
"Yinpeng Dong"
],
"summary": "Large Reasoning Models (LRMs), such as OpenAI o1 and DeepSeek-R1, have been rapidly progressing and achieving breakthrough performance on complex reasoning tasks such as mathematics and coding. However, the open-source R1 models have raised safety concerns in wide applications, such as the tendency to comply with malicious queries, which greatly impacts the utility of these powerful models in their applications. In this paper, we introduce RealSafe-R1 as safety-aligned versions of DeepSeek-R1 distilled models. To train these models, we construct a dataset of 15k safety-aware reasoning trajectories generated by DeepSeek-R1, under explicit instructions for expected refusal behavior. Both quantitative experiments and qualitative case studies demonstrate the models' improvements, which are shown in their safety guardrails against both harmful queries and jailbreak attacks. Importantly, unlike prior safety alignment efforts that often compromise reasoning performance, our method preserves the models' reasoning capabilities by maintaining the training data within the original distribution of generation. Model weights of RealSafe-R1 are open-source at https://huggingface.co/RealSafe.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2504.16027v1",
"version_id": "http://arxiv.org/abs/2504.16027v1",
"title": "Benchmarking LLM for Code Smells Detection: OpenAI GPT-4.0 vs DeepSeek-V3",
"published": "2025-04-22T16:44:39Z",
"updated": "2025-04-22T16:44:39Z",
"authors": [
"Ahmed R. Sadik",
"Siddhata Govind"
],
"summary": "Determining the most effective Large Language Model for code smell detection presents a complex challenge. This study introduces a structured methodology and evaluation matrix to tackle this issue, leveraging a curated dataset of code samples consistently annotated with known smells. The dataset spans four prominent programming languages Java, Python, JavaScript, and C++; allowing for cross language comparison. We benchmark two state of the art LLMs, OpenAI GPT 4.0 and DeepSeek-V3, using precision, recall, and F1 score as evaluation metrics. Our analysis covers three levels of detail: overall performance, category level performance, and individual code smell type performance. Additionally, we explore cost effectiveness by comparing the token based detection approach of GPT 4.0 with the pattern-matching techniques employed by DeepSeek V3. The study also includes a cost analysis relative to traditional static analysis tools such as SonarQube. The findings offer valuable guidance for practitioners in selecting an efficient, cost effective solution for automated code smell detection",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.SE"
},
{
"id": "2505.00025v2",
"version_id": "http://arxiv.org/abs/2505.00025v2",
"title": "A Method for the Architecture of a Medical Vertical Large Language Model Based on Deepseek R1",
"published": "2025-04-25T14:28:29Z",
"updated": "2025-07-22T14:26:53Z",
"authors": [
"Mingda Zhang",
"Jianglong Qin"
],
"summary": "Despite significant advances in foundation models like DeepSeek-R1 and ChatGPT, their deployment in medical settings faces critical challenges including computational requirements and professional knowledge barriers. This paper presents an efficient lightweight medical large language model architecture that systematically addresses these challenges through three-dimensional optimization: knowledge acquisition, model compression, and computational enhancement. We design a knowledge transfer pipeline from DeepSeek-R1-Distill-70B to DeepSeek-R1-Distill-7B using Low-Rank Adaptation (LoRA) for precise medical knowledge retention. Through 4-bit quantization and mixed-precision strategies, we achieve substantial model compression while preserving medical reasoning capabilities. The inference framework incorporates Flash Attention acceleration and continuous batching, complemented by specialized prompt templates for diverse medical queries. Experimental evaluation on medical benchmarks demonstrates that our approach maintains 92.1% accuracy on USMLE examinations while reducing memory consumption by 64.7% and inference latency by 12.4% compared to baseline models. This work provides a practical solution for deploying advanced language models in resource-constrained medical environments, enabling broader accessibility of AI-assisted healthcare.",
"comment": "14 pages, 1 figures",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2504.21801v2",
"version_id": "http://arxiv.org/abs/2504.21801v2",
"title": "DeepSeek-Prover-V2: Advancing Formal Mathematical Reasoning via Reinforcement Learning for Subgoal Decomposition",
"published": "2025-04-30T16:57:48Z",
"updated": "2025-07-18T08:20:23Z",
"authors": [
"Z. Z. Ren",
"Zhihong Shao",
"Junxiao Song",
"Huajian Xin",
"Haocheng Wang",
"Wanjia Zhao",
"Liyue Zhang",
"Zhe Fu",
"Qihao Zhu",
"Dejian Yang",
"Z. F. Wu",
"Zhibin Gou",
"Shirong Ma",
"Hongxuan Tang",
"Yuxuan Liu",
"Wenjun Gao",
"Daya Guo",
"Chong Ruan"
],
"summary": "We introduce DeepSeek-Prover-V2, an open-source large language model designed for formal theorem proving in Lean 4, with initialization data collected through a recursive theorem proving pipeline powered by DeepSeek-V3. The cold-start training procedure begins by prompting DeepSeek-V3 to decompose complex problems into a series of subgoals. The proofs of resolved subgoals are synthesized into a chain-of-thought process, combined with DeepSeek-V3's step-by-step reasoning, to create an initial cold start for reinforcement learning. This process enables us to integrate both informal and formal mathematical reasoning into a unified model. The resulting model, DeepSeek-Prover-V2-671B, achieves state-of-the-art performance in neural theorem proving, reaching 88.9% pass ratio on the MiniF2F-test and solving 49 out of 658 problems from PutnamBench. In addition to standard benchmarks, we introduce ProverBench, a collection of 325 formalized problems, to enrich our evaluation, including 15 selected problems from the recent AIME competitions (years 24-25). Further evaluation on these 15 AIME problems shows that the model successfully solves 6 of them. In comparison, DeepSeek-V3 solves 8 of these problems using majority voting, highlighting that the gap between formal and informal mathematical reasoning in large language models is substantially narrowing.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2505.00551v3",
"version_id": "http://arxiv.org/abs/2505.00551v3",
"title": "100 Days After DeepSeek-R1: A Survey on Replication Studies and More Directions for Reasoning Language Models",
"published": "2025-05-01T14:28:35Z",
"updated": "2025-05-15T14:16:03Z",
"authors": [
"Chong Zhang",
"Yue Deng",
"Xiang Lin",
"Bin Wang",
"Dianwen Ng",
"Hai Ye",
"Xingxuan Li",
"Yao Xiao",
"Zhanfeng Mo",
"Qi Zhang",
"Lidong Bing"
],
"summary": "The recent development of reasoning language models (RLMs) represents a novel evolution in large language models. In particular, the recent release of DeepSeek-R1 has generated widespread social impact and sparked enthusiasm in the research community for exploring the explicit reasoning paradigm of language models. However, the implementation details of the released models have not been fully open-sourced by DeepSeek, including DeepSeek-R1-Zero, DeepSeek-R1, and the distilled small models. As a result, many replication studies have emerged aiming to reproduce the strong performance achieved by DeepSeek-R1, reaching comparable performance through similar training procedures and fully open-source data resources. These works have investigated feasible strategies for supervised fine-tuning (SFT) and reinforcement learning from verifiable rewards (RLVR), focusing on data preparation and method design, yielding various valuable insights. In this report, we provide a summary of recent replication studies to inspire future research. We primarily focus on SFT and RLVR as two main directions, introducing the details for data construction, method design and training procedure of current replication studies. Moreover, we conclude key findings from the implementation details and experimental results reported by these studies, anticipating to inspire future research. We also discuss additional techniques of enhancing RLMs, highlighting the potential of expanding the application scope of these models, and discussing the challenges in development. By this survey, we aim to help researchers and developers of RLMs stay updated with the latest advancements, and seek to inspire new ideas to further enhance RLMs.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2505.02390v2",
"version_id": "http://arxiv.org/abs/2505.02390v2",
"title": "Quantitative Analysis of Performance Drop in DeepSeek Model Quantization",
"published": "2025-05-05T06:25:20Z",
"updated": "2025-06-13T06:25:10Z",
"authors": [
"Enbo Zhao",
"Yi Shen",
"Shuming Shi",
"Jieyun Huang",
"Zhihao Chen",
"Ning Wang",
"Siqi Xiao",
"Jian Zhang",
"Kai Wang",
"Shiguo Lian"
],
"summary": "Recently, there is a high demand for deploying DeepSeek-R1 and V3 locally, possibly because the official service often suffers from being busy and some organizations have data privacy concerns. While single-machine deployment offers infrastructure simplicity, the models' 671B FP8 parameter configuration exceeds the practical memory limits of a standard 8-GPU machine. Quantization is a widely used technique that helps reduce model memory consumption. However, it is unclear what the performance of DeepSeek-R1 and V3 will be after being quantized. This technical report presents the first quantitative evaluation of multi-bitwidth quantization across the complete DeepSeek model spectrum. Key findings reveal that 4-bit quantization maintains little performance degradation versus FP8 while enabling single-machine deployment on standard NVIDIA GPU devices. We further propose DQ3_K_M, a dynamic 3-bit quantization method that significantly outperforms traditional Q3_K_M variant on various benchmarks, which is also comparable with 4-bit quantization (Q4_K_M) approach in most tasks. Moreover, DQ3_K_M supports single-machine deployment configurations for both NVIDIA H100/A100 and Huawei 910B. Our implementation of DQ3\\_K\\_M is released at https://github.com/UnicomAI/DeepSeek-Eval, containing optimized 3-bit quantized variants of both DeepSeek-R1 and DeepSeek-V3.",
"comment": "This version added the results of DeepSeek-V3-0324",
"journal_ref": "",
"doi": "",
"primary": "cs.LG"
},
{
"id": "2505.09343v2",
"version_id": "http://arxiv.org/abs/2505.09343v2",
"title": "Insights into DeepSeek-V3: Scaling Challenges and Reflections on Hardware for AI Architectures",
"published": "2025-05-14T12:39:03Z",
"updated": "2025-12-23T03:00:00Z",
"authors": [
"Chenggang Zhao",
"Chengqi Deng",
"Chong Ruan",
"Damai Dai",
"Huazuo Gao",
"Jiashi Li",
"Liyue Zhang",
"Panpan Huang",
"Shangyan Zhou",
"Shirong Ma",
"Wenfeng Liang",
"Ying He",
"Yuqing Wang",
"Yuxuan Liu",
"Y. X. Wei"
],
"summary": "The rapid scaling of large language models (LLMs) has unveiled critical limitations in current hardware architectures, including constraints in memory capacity, computational efficiency, and interconnection bandwidth. DeepSeek-V3, trained on 2,048 NVIDIA H800 GPUs, demonstrates how hardware-aware model co-design can effectively address these challenges, enabling cost-efficient training and inference at scale. This paper presents an in-depth analysis of the DeepSeek-V3/R1 model architecture and its AI infrastructure, highlighting key innovations such as Multi-head Latent Attention (MLA) for enhanced memory efficiency, Mixture of Experts (MoE) architectures for optimized computation-communication trade-offs, FP8 mixed-precision training to unlock the full potential of hardware capabilities, and a Multi-Plane Network Topology to minimize cluster-level network overhead. Building on the hardware bottlenecks encountered during DeepSeek-V3's development, we engage in a broader discussion with academic and industry peers on potential future hardware directions, including precise low-precision computation units, scale-up and scale-out convergence, and innovations in low-latency communication fabrics. These insights underscore the critical role of hardware and model co-design in meeting the escalating demands of AI workloads, offering a practical blueprint for innovation in next-generation AI systems.",
"comment": "This is the author's version of the work. It is posted here for your personal use. Not for redistribution. The definitive version appeared as part of the Industry Track in Proceedings of the 52nd Annual International Symposium on Computer Architecture (ISCA '25)",
"journal_ref": "Proceedings of the 52nd Annual International Symposium on Computer Architecture (ISCA '25), June 2025",
"doi": "10.1145/3695053.3731412",
"primary": "cs.DC"
},
{
"id": "2505.12625v1",
"version_id": "http://arxiv.org/abs/2505.12625v1",
"title": "R1dacted: Investigating Local Censorship in DeepSeek's R1 Language Model",
"published": "2025-05-19T02:16:56Z",
"updated": "2025-05-19T02:16:56Z",
"authors": [
"Ali Naseh",
"Harsh Chaudhari",
"Jaechul Roh",
"Mingshi Wu",
"Alina Oprea",
"Amir Houmansadr"
],
"summary": "DeepSeek recently released R1, a high-performing large language model (LLM) optimized for reasoning tasks. Despite its efficient training pipeline, R1 achieves competitive performance, even surpassing leading reasoning models like OpenAI's o1 on several benchmarks. However, emerging reports suggest that R1 refuses to answer certain prompts related to politically sensitive topics in China. While existing LLMs often implement safeguards to avoid generating harmful or offensive outputs, R1 represents a notable shift - exhibiting censorship-like behavior on politically charged queries. In this paper, we investigate this phenomenon by first introducing a large-scale set of heavily curated prompts that get censored by R1, covering a range of politically sensitive topics, but are not censored by other models. We then conduct a comprehensive analysis of R1's censorship patterns, examining their consistency, triggers, and variations across topics, prompt phrasing, and context. Beyond English-language queries, we explore censorship behavior in other languages. We also investigate the transferability of censorship to models distilled from the R1 language model. Finally, we propose techniques for bypassing or removing this censorship. Our findings reveal possible additional censorship integration likely shaped by design choices during training or alignment, raising concerns about transparency, bias, and governance in language model deployment.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2505.17112v1",
"version_id": "http://arxiv.org/abs/2505.17112v1",
"title": "Cultural Value Alignment in Large Language Models: A Prompt-based Analysis of Schwartz Values in Gemini, ChatGPT, and DeepSeek",
"published": "2025-05-21T14:03:19Z",
"updated": "2025-05-21T14:03:19Z",
"authors": [
"Robin Segerer"
],
"summary": "This study examines cultural value alignment in large language models (LLMs) by analyzing how Gemini, ChatGPT, and DeepSeek prioritize values from Schwartz's value framework. Using the 40-item Portrait Values Questionnaire, we assessed whether DeepSeek, trained on Chinese-language data, exhibits distinct value preferences compared to Western models. Results of a Bayesian ordinal regression model show that self-transcendence values (e.g., benevolence, universalism) were highly prioritized across all models, reflecting a general LLM tendency to emphasize prosocial values. However, DeepSeek uniquely downplayed self-enhancement values (e.g., power, achievement) compared to ChatGPT and Gemini, aligning with collectivist cultural tendencies. These findings suggest that LLMs reflect culturally situated biases rather than a universal ethical framework. To address value asymmetries in LLMs, we propose multi-perspective reasoning, self-reflective feedback, and dynamic contextualization. This study contributes to discussions on AI fairness, cultural neutrality, and the need for pluralistic AI alignment frameworks that integrate diverse moral perspectives.",
"comment": "15 pages, 1 table, 1 figure",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2505.17702v3",
"version_id": "http://arxiv.org/abs/2505.17702v3",
"title": "Seek-CAD: A Self-refined Generative Modeling for 3D Parametric CAD Using Local Inference via DeepSeek",
"published": "2025-05-23T10:11:19Z",
"updated": "2026-02-28T15:25:47Z",
"authors": [
"Xueyang Li",
"Jiahao Li",
"Yu Song",
"Yunzhong Lou",
"Xiangdong Zhou"
],
"summary": "The advent of Computer-Aided Design (CAD) generative modeling will significantly transform the design of industrial products. The recent research endeavor has extended into the realm of Large Language Models (LLMs). In contrast to fine-tuning methods, training-free approaches typically utilize the advanced closed-source LLMs, thereby offering enhanced flexibility and efficiency in the development of AI agents for generating CAD parametric models. However, the substantial cost and limitations of local deployment of the top-tier closed-source LLMs pose challenges in practical applications. The Seek-CAD is the pioneer exploration of locally deployed open-source inference LLM DeepSeek-R1 for CAD parametric model generation with a training-free methodology. This study is the first investigation to incorporate both visual and Chain-of-Thought (CoT) feedback within the self-refinement mechanism for generating CAD models. Specifically, the initial generated parametric CAD model is rendered into a sequence of step-wise perspective images, which are subsequently processed by a Vision Language Model (VLM) alongside the corresponding CoTs derived from DeepSeek-R1 to assess the CAD model generation. Then, the feedback is utilized by DeepSeek-R1 to refine the initial generated model for the next round of generation. Moreover, we present an innovative 3D CAD model dataset structured around the SSR (Sketch, Sketch-based feature, and Refinements) triple design paradigm. This dataset encompasses a wide range of CAD commands, thereby aligning effectively with industrial application requirements and proving suitable for the generation of LLMs. Extensive experiments validate the effectiveness of Seek-CAD under various metrics.",
"comment": "Accepted to ICLR 2026. The datatset has been released publicly and can be acessed in https://github.com/Sunny-Hack/Seek-CAD",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2505.18059v1",
"version_id": "http://arxiv.org/abs/2505.18059v1",
"title": "Assessing the performance of 8 AI chatbots in bibliographic reference retrieval: Grok and DeepSeek outperform ChatGPT, but none are fully accurate",
"published": "2025-05-23T16:07:14Z",
"updated": "2025-05-23T16:07:14Z",
"authors": [
"Álvaro Cabezas-Clavijo",
"Pavel Sidorenko-Bautista"
],
"summary": "This study analyzes the performance of eight generative artificial intelligence chatbots -- ChatGPT, Claude, Copilot, DeepSeek, Gemini, Grok, Le Chat, and Perplexity -- in their free versions, in the task of generating academic bibliographic references within the university context. A total of 400 references were evaluated across the five major areas of knowledge (Health, Engineering, Experimental Sciences, Social Sciences, and Humanities), based on a standardized prompt. Each reference was assessed according to five key components (authorship, year, title, source, and location), along with document type, publication age, and error count. The results show that only 26.5% of the references were fully correct, 33.8% partially correct, and 39.8% were either erroneous or entirely fabricated. Grok and DeepSeek stood out as the only chatbots that did not generate false references, while Copilot, Perplexity, and Claude exhibited the highest hallucination rates. Furthermore, the chatbots showed a greater tendency to generate book references over journal articles, although the latter had a significantly higher fabrication rate. A high degree of overlap was also detected among the sources provided by several models, particularly between DeepSeek, Grok, Gemini, and ChatGPT. These findings reveal structural limitations in current AI models, highlight the risks of uncritical use by students, and underscore the need to strengthen information and critical literacy regarding the use of AI tools in higher education.",
"comment": "Journal of Data and Information Science (2026)",
"journal_ref": "",
"doi": "10.1515/jdis-2025-0326",
"primary": "cs.IR"
},
{
"id": "2506.00312v1",
"version_id": "http://arxiv.org/abs/2506.00312v1",
"title": "An evaluation of LLMs for generating movie reviews: GPT-4o, Gemini-2.0 and DeepSeek-V3",
"published": "2025-05-30T23:45:53Z",
"updated": "2025-05-30T23:45:53Z",
"authors": [
"Brendan Sands",
"Yining Wang",
"Chenhao Xu",
"Yuxuan Zhou",
"Lai Wei",
"Rohitash Chandra"
],
"summary": "Large language models (LLMs) have been prominent in various tasks, including text generation and summarisation. The applicability of LLMs to the generation of product reviews is gaining momentum, paving the way for the generation of movie reviews. In this study, we propose a framework that generates movie reviews using three LLMs (GPT-4o, DeepSeek-V3, and Gemini-2.0), and evaluate their performance by comparing the generated outputs with IMDb user reviews. We use movie subtitles and screenplays as input to the LLMs and investigate how they affect the quality of reviews generated. We review the LLM-based movie reviews in terms of vocabulary, sentiment polarity, similarity, and thematic consistency in comparison to IMDB user reviews. The results demonstrate that LLMs are capable of generating syntactically fluent and structurally complete movie reviews. Nevertheless, there is still a noticeable gap in emotional richness and stylistic coherence between LLM-generated and IMDb reviews, suggesting that further refinement is needed to improve the overall quality of movie review generation. We provided a survey-based analysis where participants were told to distinguish between LLM and IMDb user reviews. The results show that LLM-generated reviews are difficult to distinguish from IMDB user reviews. We found that DeepSeek-V3 produced the most balanced reviews, closely matching IMDb reviews. GPT-4o overemphasised positive emotions, while Gemini-2.0 captured negative emotions better but showed excessive emotional intensity.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2507.19493v1",
"version_id": "http://arxiv.org/abs/2507.19493v1",
"title": "From Bench to Bedside: A DeepSeek-Powered AI System for Automated Chest Radiograph Interpretation in Clinical Practice",
"published": "2025-05-31T06:13:50Z",
"updated": "2025-05-31T06:13:50Z",
"authors": [
"Yaowei Bai",
"Ruiheng Zhang",
"Yu Lei",
"Jingfeng Yao",
"Shuguang Ju",
"Chaoyang Wang",
"Wei Yao",
"Yiwan Guo",
"Guilin Zhang",
"Chao Wan",
"Qian Yuan",
"Xuhua Duan",
"Xinggang Wang",
"Tao Sun",
"Yongchao Xu",
"Chuansheng Zheng",
"Huangxuan Zhao",
"Bo Du"
],
"summary": "A global shortage of radiologists has been exacerbated by the significant volume of chest X-ray workloads, particularly in primary care. Although multimodal large language models show promise, existing evaluations predominantly rely on automated metrics or retrospective analyses, lacking rigorous prospective clinical validation. Janus-Pro-CXR (1B), a chest X-ray interpretation system based on DeepSeek Janus-Pro model, was developed and rigorously validated through a multicenter prospective trial (NCT06874647). Our system outperforms state-of-the-art X-ray report generation models in automated report generation, surpassing even larger-scale models including ChatGPT 4o (200B parameters), while demonstrating robust detection of eight clinically critical radiographic findings (area under the curve, AUC > 0.8). Retrospective evaluation confirms significantly higher report accuracy than Janus-Pro and ChatGPT 4o. In prospective clinical deployment, AI assistance significantly improved report quality scores (4.37 vs. 4.11, P < 0.001), reduced interpretation time by 18.5% (P < 0.001), and was preferred by a majority of experts (3 out of 5) in 52.7% of cases. Through lightweight architecture and domain-specific optimization, Janus-Pro-CXR improves diagnostic reliability and workflow efficiency, particularly in resource-constrained settings. The model architecture and implementation framework will be open-sourced to facilitate the clinical translation of AI-assisted radiology solutions.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.HC"
},
{
"id": "2506.01257v1",
"version_id": "http://arxiv.org/abs/2506.01257v1",
"title": "DeepSeek in Healthcare: A Survey of Capabilities, Risks, and Clinical Applications of Open-Source Large Language Models",
"published": "2025-06-02T02:17:04Z",
"updated": "2025-06-02T02:17:04Z",
"authors": [
"Jiancheng Ye",
"Sophie Bronstein",
"Jiarui Hai",
"Malak Abu Hashish"
],
"summary": "DeepSeek-R1 is a cutting-edge open-source large language model (LLM) developed by DeepSeek, showcasing advanced reasoning capabilities through a hybrid architecture that integrates mixture of experts (MoE), chain of thought (CoT) reasoning, and reinforcement learning. Released under the permissive MIT license, DeepSeek-R1 offers a transparent and cost-effective alternative to proprietary models like GPT-4o and Claude-3 Opus; it excels in structured problem-solving domains such as mathematics, healthcare diagnostics, code generation, and pharmaceutical research. The model demonstrates competitive performance on benchmarks like the United States Medical Licensing Examination (USMLE) and American Invitational Mathematics Examination (AIME), with strong results in pediatric and ophthalmologic clinical decision support tasks. Its architecture enables efficient inference while preserving reasoning depth, making it suitable for deployment in resource-constrained settings. However, DeepSeek-R1 also exhibits increased vulnerability to bias, misinformation, adversarial manipulation, and safety failures - especially in multilingual and ethically sensitive contexts. This survey highlights the model's strengths, including interpretability, scalability, and adaptability, alongside its limitations in general language fluency and safety alignment. Future research priorities include improving bias mitigation, natural language comprehension, domain-specific validation, and regulatory compliance. Overall, DeepSeek-R1 represents a major advance in open, scalable AI, underscoring the need for collaborative governance to ensure responsible and equitable deployment.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2506.01814v1",
"version_id": "http://arxiv.org/abs/2506.01814v1",
"title": "Analysis of LLM Bias (Chinese Propaganda & Anti-US Sentiment) in DeepSeek-R1 vs. ChatGPT o3-mini-high",
"published": "2025-06-02T15:54:06Z",
"updated": "2025-06-02T15:54:06Z",
"authors": [
"PeiHsuan Huang",
"ZihWei Lin",
"Simon Imbot",
"WenCheng Fu",
"Ethan Tu"
],
"summary": "Large language models (LLMs) increasingly shape public understanding and civic decisions, yet their ideological neutrality is a growing concern. While existing research has explored various forms of LLM bias, a direct, cross-lingual comparison of models with differing geopolitical alignments-specifically a PRC-system model versus a non-PRC counterpart-has been lacking. This study addresses this gap by systematically evaluating DeepSeek-R1 (PRC-aligned) against ChatGPT o3-mini-high (non-PRC) for Chinese-state propaganda and anti-U.S. sentiment. We developed a novel corpus of 1,200 de-contextualized, reasoning-oriented questions derived from Chinese-language news, presented in Simplified Chinese, Traditional Chinese, and English. Answers from both models (7,200 total) were assessed using a hybrid evaluation pipeline combining rubric-guided GPT-4o scoring with human annotation. Our findings reveal significant model-level and language-dependent biases. DeepSeek-R1 consistently exhibited substantially higher proportions of both propaganda and anti-U.S. bias compared to ChatGPT o3-mini-high, which remained largely free of anti-U.S. sentiment and showed lower propaganda levels. For DeepSeek-R1, Simplified Chinese queries elicited the highest bias rates; these diminished in Traditional Chinese and were nearly absent in English. Notably, DeepSeek-R1 occasionally responded in Simplified Chinese to Traditional Chinese queries and amplified existing PRC-aligned terms in its Chinese answers, demonstrating an \"invisible loudspeaker\" effect. Furthermore, such biases were not confined to overtly political topics but also permeated cultural and lifestyle content, particularly in DeepSeek-R1.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2506.02523v1",
"version_id": "http://arxiv.org/abs/2506.02523v1",
"title": "Hardware-Centric Analysis of DeepSeek's Multi-Head Latent Attention",
"published": "2025-06-03T06:53:04Z",
"updated": "2025-06-03T06:53:04Z",
"authors": [
"Robin Geens",
"Marian Verhelst"
],
"summary": "Multi-Head Latent Attention (MLA), introduced in DeepSeek-V2, improves the efficiency of large language models by projecting query, key, and value tensors into a compact latent space. This architectural change reduces the KV-cache size and significantly lowers memory bandwidth demands, particularly in the autoregressive decode phase. This letter presents the first hardware-centric analysis of MLA, comparing it to conventional Multi-Head Attention (MHA) and evaluating its implications for accelerator performance. We identify two alternative execution schemes of MLA--reusing, resp. recomputing latent projection matrices--which offer distinct trade-offs between compute and memory access. Using the Stream design space exploration framework, we model their throughput and energy cost across a range of hardware platforms and find that MLA can shift attention workloads toward the compute-bound regime. Our results show that MLA not only reduces bandwidth usage but also enables adaptable execution strategies aligned with hardware constraints. Compared to MHA, it provides more stable and efficient performance, particularly on bandwidth-limited hardware platforms. These findings emphasize MLA's relevance as a co-design opportunity for future AI accelerators.",
"comment": "",
"journal_ref": "",
"doi": "10.1049/ell2.70504",
"primary": "cs.AR"
},
{
"id": "2506.05921v1",
"version_id": "http://arxiv.org/abs/2506.05921v1",
"title": "Multi-Modal Large Models Based Beam Prediction: An Example Empowered by DeepSeek",
"published": "2025-06-06T09:43:24Z",
"updated": "2025-06-06T09:43:24Z",
"authors": [
"Yizhu Zhao",
"Li Yu",
"Lianzheng Shi",
"Jianhua Zhang",
"Guangyi Liu"
],
"summary": "Beam prediction is an effective approach to reduce training overhead in massive multiple-input multiple-output (MIMO) systems. However, existing beam prediction models still exhibit limited generalization ability in diverse scenarios, which remains a critical challenge. In this paper, we propose MLM-BP, a beam prediction framework based on the multi-modal large model released by DeepSeek, with full consideration of multi-modal environmental information. Specifically, the distribution of scatterers that impact the optimal beam is captured by the sensing devices. Then positions are tokenized to generate text-based representations, and multi-view images are processed by an image encoder, which is fine-tuned with low-rank adaptation (LoRA), to extract environmental embeddings. Finally, these embeddings are fed into the large model, and an output projection module is designed to determine the optimal beam index. Simulation results show that MLM-BP achieves 98.1% Top-1 accuracy on the simulation dataset. Additionally, it demonstrates few-shot generalization on a real-world dataset, achieving 72.7% Top-1 accuracy and 92.4% Top-3 accuracy with only 30% of the dataset, outperforming the existing small models by over 15%.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "eess.SP"
},
{
"id": "2506.21562v2",
"version_id": "http://arxiv.org/abs/2506.21562v2",
"title": "FloorPlan-DeepSeek (FPDS): A multimodal approach to floorplan generation using vector-based next room prediction",
"published": "2025-06-12T04:33:27Z",
"updated": "2025-08-02T18:27:22Z",
"authors": [
"Jun Yin",
"Pengyu Zeng",
"Jing Zhong",
"Peilin Li",
"Miao Zhang",
"Ran Luo",
"Shuai Lu"
],
"summary": "In the architectural design process, floor plan generation is inherently progressive and iterative. However, existing generative models for floor plans are predominantly end-to-end generation that produce an entire pixel-based layout in a single pass. This paradigm is often incompatible with the incremental workflows observed in real-world architectural practice. To address this issue, we draw inspiration from the autoregressive 'next token prediction' mechanism commonly used in large language models, and propose a novel 'next room prediction' paradigm tailored to architectural floor plan modeling. Experimental evaluation indicates that FPDS demonstrates competitive performance in comparison to diffusion models and Tell2Design in the text-to-floorplan task, indicating its potential applicability in supporting future intelligent architectural design.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2506.12349v1",
"version_id": "http://arxiv.org/abs/2506.12349v1",
"title": "Information Suppression in Large Language Models: Auditing, Quantifying, and Characterizing Censorship in DeepSeek",
"published": "2025-06-14T05:01:50Z",
"updated": "2025-06-14T05:01:50Z",
"authors": [
"Peiran Qiu",
"Siyi Zhou",
"Emilio Ferrara"
],
"summary": "This study examines information suppression mechanisms in DeepSeek, an open-source large language model (LLM) developed in China. We propose an auditing framework and use it to analyze the model's responses to 646 politically sensitive prompts by comparing its final output with intermediate chain-of-thought (CoT) reasoning. Our audit unveils evidence of semantic-level information suppression in DeepSeek: sensitive content often appears within the model's internal reasoning but is omitted or rephrased in the final output. Specifically, DeepSeek suppresses references to transparency, government accountability, and civic mobilization, while occasionally amplifying language aligned with state propaganda. This study underscores the need for systematic auditing of alignment, content moderation, information suppression, and censorship practices implemented into widely-adopted AI models, to ensure transparency, accountability, and equitable access to unbiased information obtained by means of these systems.",
"comment": "",
"journal_ref": "Inf. Sci. 724, C (Jan 2026)",
"doi": "10.1016/j.ins.2025.122702",
"primary": "cs.CY"
},
{
"id": "2506.14104v1",
"version_id": "http://arxiv.org/abs/2506.14104v1",
"title": "Innovating China's Intangible Cultural Heritage with DeepSeek + MidJourney: The Case of Yangliuqing theme Woodblock Prints",
"published": "2025-06-17T01:47:17Z",
"updated": "2025-06-17T01:47:17Z",
"authors": [
"RuiKun Yang",
"ZhongLiang Wei",
"Longdi Xian"
],
"summary": "Yangliuqing woodblock prints, a cornerstone of China's intangible cultural heritage, are celebrated for their intricate designs and vibrant colors. However, preserving these traditional art forms while fostering innovation presents significant challenges. This study explores the DeepSeek + MidJourney approach to generating creative, themed Yangliuqing woodblock prints focused on the fight against COVID-19 and depicting joyous winners. Using Fréchet Inception Distance (FID) scores for evaluation, the method that combined DeepSeek-generated thematic prompts, MidJourney-generated thematic images, original Yangliuqing prints, and DeepSeek-generated key prompts in MidJourney-generated outputs achieved the lowest mean FID score (150.2) with minimal variability (σ = 4.9). Additionally, feedback from 62 participants, collected via questionnaires, confirmed that this hybrid approach produced the most representative results. Moreover, the questionnaire data revealed that participants demonstrated the highest willingness to promote traditional culture and the strongest interest in consuming the AI-generated images produced through this method. These findings underscore the effectiveness of an innovative approach that seamlessly blends traditional artistic elements with modern AI-driven creativity, ensuring both cultural preservation and contemporary relevance.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.GR"
},
{
"id": "2506.18501v3",
"version_id": "http://arxiv.org/abs/2506.18501v3",
"title": "Comparative Evaluation of ChatGPT and DeepSeek Across Key NLP Tasks: Strengths, Weaknesses, and Domain-Specific Performance",
"published": "2025-06-23T10:52:54Z",
"updated": "2025-07-05T12:37:20Z",
"authors": [
"Wael Etaiwi",
"Bushra Alhijawi"
],
"summary": "The increasing use of large language models (LLMs) in natural language processing (NLP) tasks has sparked significant interest in evaluating their effectiveness across diverse applications. While models like ChatGPT and DeepSeek have shown strong results in many NLP domains, a comprehensive evaluation is needed to understand their strengths, weaknesses, and domain-specific abilities. This is critical as these models are applied to various tasks, from sentiment analysis to more nuanced tasks like textual entailment and translation. This study aims to evaluate ChatGPT and DeepSeek across five key NLP tasks: sentiment analysis, topic classification, text summarization, machine translation, and textual entailment. A structured experimental protocol is used to ensure fairness and minimize variability. Both models are tested with identical, neutral prompts and evaluated on two benchmark datasets per task, covering domains like news, reviews, and formal/informal texts. The results show that DeepSeek excels in classification stability and logical reasoning, while ChatGPT performs better in tasks requiring nuanced understanding and flexibility. These findings provide valuable insights for selecting the appropriate LLM based on task requirements.",
"comment": "",
"journal_ref": "",
"doi": "10.1016/j.array.2025.100478",
"primary": "cs.CL"
},
{
"id": "2506.18543v2",
"version_id": "http://arxiv.org/abs/2506.18543v2",
"title": "SoK: A Comprehensive Security Analysis of Jailbreak Resilience in GPT and DeepSeek Models",
"published": "2025-06-23T11:53:31Z",
"updated": "2026-05-25T10:15:56Z",
"authors": [
"Xiaodong Wu",
"Xiangman Li",
"Qi Li",
"Lingshuang Liu",
"Jianbing Ni"
],
"summary": "The rapid proliferation of Large Language Models (LLMs) has heightened concerns regarding their exposure to jailbreak attacks, which craft adversarial inputs designed to elicit unsafe content. Although proprietary models such as GPT-4 have been extensively evaluated, the robustness of emerging open-source systems like DeepSeek remains insufficiently examined, despite their growing use in LLM applications. In this paper, we conduct the first comprehensive jailbreak analysis of the DeepSeek model family, comparing it with GPT-3.5 and GPT-4 through the HarmBench benchmark. We investigate seven representative attack methods across 510 harmful behaviors, organized along both functional and semantic dimensions. Findings indicate that DeepSeek provides partial resilience against optimization-driven attacks such as TAP-T, but also results in greater susceptibility to prompt-based and manually engineered adversarial inputs. In contrast, GPT-4 Turbo demonstrates more robust and consistent safety alignment across a wide range of behaviors, likely due to stronger safety optimization and reinforcement learning from human feedback. In addition, fine-grained behavioral analysis and case studies reveal that DeepSeek often fails to consistently apply safety constraints to adversarial prompts, leading to uneven refusal behaviors. Overall, our results highlight an inherent trade-off between model efficiency and alignment generalization, underscoring the importance of targeted safety tuning and robust alignment strategies to ensure secure deployment of open-source LLMs.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CR"
},
{
"id": "2506.23128v1",
"version_id": "http://arxiv.org/abs/2506.23128v1",
"title": "Are Large Language Models Capable of Deep Relational Reasoning? Insights from DeepSeek-R1 and Benchmark Comparisons",
"published": "2025-06-29T07:37:49Z",
"updated": "2025-06-29T07:37:49Z",
"authors": [
"Chi Chiu So",
"Yueyue Sun",
"Jun-Min Wang",
"Siu Pang Yung",
"Anthony Wai Keung Loh",
"Chun Pong Chau"
],
"summary": "How far are Large Language Models (LLMs) in performing deep relational reasoning? In this paper, we evaluate and compare the reasoning capabilities of three cutting-edge LLMs, namely, DeepSeek-R1, DeepSeek-V3 and GPT-4o, through a suite of carefully designed benchmark tasks in family tree and general graph reasoning. Our experiments reveal that DeepSeek-R1 consistently achieves the highest F1-scores across multiple tasks and problem sizes, demonstrating strong aptitude in logical deduction and relational inference. However, all evaluated models, including DeepSeek-R1, struggle significantly as problem complexity increases, largely due to token length limitations and incomplete output structures. A detailed analysis of DeepSeek-R1's long Chain-of-Thought responses uncovers its unique planning and verification strategies, but also highlights instances of incoherent or incomplete reasoning, calling attention to the need for deeper scrutiny into LLMs' internal inference dynamics. We further discuss key directions for future work, including the role of multimodal reasoning and the systematic examination of reasoning failures. Our findings provide both empirical insights and theoretical implications for advancing LLMs' reasoning abilities, particularly in tasks that demand structured, multi-step logical inference. Our code repository will be publicly available at https://github.com/kelvinhkcs/Deep-Relational-Reasoning.",
"comment": "10 pages, 0 figures, accepted by 2025 IEEE international conference on artificial intelligence testing (AITest)",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2507.08621v2",
"version_id": "http://arxiv.org/abs/2507.08621v2",
"title": "A comprehensive study of LLM-based argument classification: from LLAMA through GPT-4o to Deepseek-R1",
"published": "2025-07-11T14:23:40Z",
"updated": "2025-07-24T11:49:06Z",
"authors": [
"Marcin Pietroń",
"Rafał Olszowski",
"Jakub Gomułka",
"Filip Gampel",
"Andrzej Tomski"
],
"summary": "Argument mining (AM) is an interdisciplinary research field that integrates insights from logic, philosophy, linguistics, rhetoric, law, psychology, and computer science. It involves the automatic identification and extraction of argumentative components, such as premises and claims, and the detection of relationships between them, such as support, attack, or neutrality. Recently, the field has advanced significantly, especially with the advent of large language models (LLMs), which have enhanced the efficiency of analyzing and extracting argument semantics compared to traditional methods and other deep learning models. There are many benchmarks for testing and verifying the quality of LLM, but there is still a lack of research and results on the operation of these models in publicly available argument classification databases. This paper presents a study of a selection of LLM's, using diverse datasets such as Args.me and UKP. The models tested include versions of GPT, Llama, and DeepSeek, along with reasoning-enhanced variants incorporating the Chain-of-Thoughts algorithm. The results indicate that ChatGPT-4o outperforms the others in the argument classification benchmarks. In case of models incorporated with reasoning capabilities, the Deepseek-R1 shows its superiority. However, despite their superiority, GPT-4o and Deepseek-R1 still make errors. The most common errors are discussed for all models. To our knowledge, the presented work is the first broader analysis of the mentioned datasets using LLM and prompt algorithms. The work also shows some weaknesses of known prompt algorithms in argument analysis, while indicating directions for their improvement. The added value of the work is the in-depth analysis of the available argument datasets and the demonstration of their shortcomings.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2507.09955v1",
"version_id": "http://arxiv.org/abs/2507.09955v1",
"title": "DeepSeek: Paradigm Shifts and Technical Evolution in Large AI Models",
"published": "2025-07-14T06:10:30Z",
"updated": "2025-07-14T06:10:30Z",
"authors": [
"Luolin Xiong",
"Haofen Wang",
"Xi Chen",
"Lu Sheng",
"Yun Xiong",
"Jingping Liu",
"Yanghua Xiao",
"Huajun Chen",
"Qing-Long Han",
"Yang Tang"
],
"summary": "DeepSeek, a Chinese Artificial Intelligence (AI) startup, has released their V3 and R1 series models, which attracted global attention due to their low cost, high performance, and open-source advantages. This paper begins by reviewing the evolution of large AI models focusing on paradigm shifts, the mainstream Large Language Model (LLM) paradigm, and the DeepSeek paradigm. Subsequently, the paper highlights novel algorithms introduced by DeepSeek, including Multi-head Latent Attention (MLA), Mixture-of-Experts (MoE), Multi-Token Prediction (MTP), and Group Relative Policy Optimization (GRPO). The paper then explores DeepSeek engineering breakthroughs in LLM scaling, training, inference, and system-level optimization architecture. Moreover, the impact of DeepSeek models on the competitive AI landscape is analyzed, comparing them to mainstream LLMs across various fields. Finally, the paper reflects on the insights gained from DeepSeek innovations and discusses future trends in the technical and engineering development of large AI models, particularly in data, training, and reasoning.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2507.17944v1",
"version_id": "http://arxiv.org/abs/2507.17944v1",
"title": "Evaluating the Performance of AI Text Detectors, Few-Shot and Chain-of-Thought Prompting Using DeepSeek Generated Text",
"published": "2025-07-23T21:26:33Z",
"updated": "2025-07-23T21:26:33Z",
"authors": [
"Hulayyil Alshammari",
"Praveen Rao"
],
"summary": "Large language models (LLMs) have rapidly transformed the creation of written materials. LLMs have led to questions about writing integrity, thereby driving the creation of artificial intelligence (AI) detection technologies. Adversarial attacks, such as standard and humanized paraphrasing, inhibit detectors' ability to detect machine-generated text. Previous studies have mainly focused on ChatGPT and other well-known LLMs and have shown varying accuracy across detectors. However, there is a clear gap in the literature about DeepSeek, a recently published LLM. Therefore, in this work, we investigate whether six generally accessible AI detection tools -- AI Text Classifier, Content Detector AI, Copyleaks, QuillBot, GPT-2, and GPTZero -- can consistently recognize text generated by DeepSeek. The detectors were exposed to the aforementioned adversarial attacks. We also considered DeepSeek as a detector by performing few-shot prompting and chain-of-thought reasoning (CoT) for classifying AI and human-written text. We collected 49 human-authored question-answer pairs from before the LLM era and generated matching responses using DeepSeek-v3, producing 49 AI-generated samples. Then, we applied adversarial techniques such as paraphrasing and humanizing to add 196 more samples. These were used to challenge detector robustness and assess accuracy impact. While QuillBot and Copyleaks showed near-perfect performance on original and paraphrased DeepSeek text, others -- particularly AI Text Classifier and GPT-2 -- showed inconsistent results. The most effective attack was humanization, reducing accuracy to 71% for Copyleaks, 58% for QuillBot, and 52% for GPTZero. Few-shot and CoT prompting showed high accuracy, with the best five-shot result misclassifying only one of 49 samples (AI recall 96%, human recall 100%).",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2508.10173v1",
"version_id": "http://arxiv.org/abs/2508.10173v1",
"title": "Benchmark-Driven Selection of AI: Evidence from DeepSeek-R1",
"published": "2025-08-13T20:15:20Z",
"updated": "2025-08-13T20:15:20Z",
"authors": [
"Petr Spelda",
"Vit Stritecky"
],
"summary": "Evaluation of reasoning language models gained importance after it was observed that they can combine their existing capabilities into novel traces of intermediate steps before task completion and that the traces can sometimes help them to generalize better than past models. As reasoning becomes the next scaling dimension of large language models, careful study of their capabilities in critical tasks is needed. We show that better performance is not always caused by test-time algorithmic improvements or model sizes but also by using impactful benchmarks as curricula for learning. We call this benchmark-driven selection of AI and show its effects on DeepSeek-R1 using our sequential decision-making problem from Humanity's Last Exam. Steering development of AI by impactful benchmarks trades evaluation for learning and makes novelty of test tasks key for measuring generalization capabilities of reasoning models. Consequently, some benchmarks could be seen as curricula for training rather than unseen test sets.",
"comment": "17 pages, 5 figures, 2 tables",
"journal_ref": "",
"doi": "",
"primary": "cs.LG"
},
{
"id": "2508.21377v1",
"version_id": "http://arxiv.org/abs/2508.21377v1",
"title": "Challenges and Applications of Large Language Models: A Comparison of GPT and DeepSeek family of models",
"published": "2025-08-29T07:41:04Z",
"updated": "2025-08-29T07:41:04Z",
"authors": [
"Shubham Sharma",
"Sneha Tuli",
"Narendra Badam"
],
"summary": "Large Language Models (LLMs) are transforming AI across industries, but their development and deployment remain complex. This survey reviews 16 key challenges in building and using LLMs and examines how these challenges are addressed by two state-of-the-art models with unique approaches: OpenAI's closed source GPT-4o (May 2024 update) and DeepSeek-V3-0324 (March 2025), a large open source Mixture-of-Experts model. Through this comparison, we showcase the trade-offs between closed source models (robust safety, fine-tuned reliability) and open source models (efficiency, adaptability). We also explore LLM applications across different domains (from chatbots and coding tools to healthcare and education), highlighting which model attributes are best suited for each use case. This article aims to guide AI researchers, developers, and decision-makers in understanding current LLM capabilities, limitations, and best practices.",
"comment": "18 pages, 7 figures",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2509.02036v1",
"version_id": "http://arxiv.org/abs/2509.02036v1",
"title": "DeepSeek performs better than other Large Language Models in Dental Cases",
"published": "2025-09-02T07:26:20Z",
"updated": "2025-09-02T07:26:20Z",
"authors": [
"Hexian Zhang",
"Xinyu Yan",
"Yanqi Yang",
"Lijian Jin",
"Ping Yang",
"Junwen Wang"
],
"summary": "Large language models (LLMs) hold transformative potential in healthcare, yet their capacity to interpret longitudinal patient narratives remains inadequately explored. Dentistry, with its rich repository of structured clinical data, presents a unique opportunity to rigorously assess LLMs' reasoning abilities. While several commercial LLMs already exist, DeepSeek, a model that gained significant attention earlier this year, has also joined the competition. This study evaluated four state-of-the-art LLMs (GPT-4o, Gemini 2.0 Flash, Copilot, and DeepSeek V3) on their ability to analyze longitudinal dental case vignettes through open-ended clinical tasks. Using 34 standardized longitudinal periodontal cases (comprising 258 question-answer pairs), we assessed model performance via automated metrics and blinded evaluations by licensed dentists. DeepSeek emerged as the top performer, demonstrating superior faithfulness (median score = 0.528 vs. 0.367-0.457) and higher expert ratings (median = 4.5/5 vs. 4.0/5), without significantly compromising readability. Our study positions DeepSeek as the leading LLM for case analysis, endorses its integration as an adjunct tool in both medical education and research, and highlights its potential as a domain-specific agent.",
"comment": "Abstract word count: 171; Total word count: 3130; Total number of tables: 2; Total number of figures: 3; Number of references: 32",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2509.19346v1",
"version_id": "http://arxiv.org/abs/2509.19346v1",
"title": "Benchmarking ChatGPT and DeepSeek in April 2025: A Novel Dual Perspective Sentiment Analysis Using Lexicon-Based and Deep Learning Approaches",
"published": "2025-09-16T20:58:10Z",
"updated": "2025-09-16T20:58:10Z",
"authors": [
"Maryam Mahdi Alhusseini",
"Mohammad-Reza Feizi-Derakhshi"
],
"summary": "This study presents a novel dual-perspective approach to analyzing user reviews for ChatGPT and DeepSeek on the Google Play Store, integrating lexicon-based sentiment analysis (TextBlob) with deep learning classification models, including Convolutional Neural Networks (CNN) and Bidirectional Long Short Term Memory (Bi LSTM) Networks. Unlike prior research, which focuses on either lexicon-based strategies or predictive deep learning models in isolation, this study conducts an extensive investigation into user satisfaction with Large Language Model (LLM) based applications. A Dataset of 4,000 authentic user reviews was collected, which were carefully preprocessed and subjected to oversampling to achieve balanced classes. The balanced test set of 1,700 Reviews were used for model testing. Results from the experiments reveal that ChatGPT received significantly more positive sentiment than DeepSeek. Furthermore, deep learning based classification demonstrated superior performance over lexicon analysis, with CNN outperforming Bi-LSTM by achieving 96.41 percent accuracy and near perfect classification of negative reviews, alongside high F1-scores for neutral and positive sentiments. This research sets a new methodological standard for measuring sentiment in LLM-based applications and provides practical insights for developers and researchers seeking to improve user-centric AI system design.",
"comment": "17 pages, 21 figures",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2509.23676v1",
"version_id": "http://arxiv.org/abs/2509.23676v1",
"title": "From Reasoning to Answer: Empirical, Attention-Based and Mechanistic Insights into Distilled DeepSeek R1 Models",
"published": "2025-09-28T06:32:21Z",
"updated": "2025-09-28T06:32:21Z",
"authors": [
"Jue Zhang",
"Qingwei Lin",
"Saravan Rajmohan",
"Dongmei Zhang"
],
"summary": "Large Reasoning Models (LRMs) generate explicit reasoning traces alongside final answers, yet the extent to which these traces influence answer generation remains unclear. In this work, we conduct a three-stage investigation into the interplay between reasoning and answer generation in three distilled DeepSeek R1 models. First, through empirical evaluation, we demonstrate that including explicit reasoning consistently improves answer quality across diverse domains. Second, attention analysis reveals that answer tokens attend substantially to reasoning tokens, with certain mid-layer Reasoning-Focus Heads (RFHs) closely tracking the reasoning trajectory, including self-reflective cues. Third, we apply mechanistic interventions using activation patching to assess the dependence of answer tokens on reasoning activations. Our results show that perturbations to key reasoning tokens can reliably alter the final answers, confirming a directional and functional flow of information from reasoning to answer. These findings deepen our understanding of how LRMs leverage reasoning tokens for answer generation, highlighting the functional role of intermediate reasoning in shaping model outputs. Our data and code are publicly available at \\href{https://aka.ms/R2A-code}{this URL}.",
"comment": "Accepted by EMNLP'25 (Main)",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2510.15009v1",
"version_id": "http://arxiv.org/abs/2510.15009v1",
"title": "Can generative AI figure out figurative language? The influence of idioms on essay scoring by ChatGPT, Gemini, and Deepseek",
"published": "2025-10-16T08:14:52Z",
"updated": "2025-10-16T08:14:52Z",
"authors": [
"Enis Oğuz"
],
"summary": "The developments in Generative AI technologies have paved the way for numerous innovations in different fields. Recently, Generative AI has been proposed as a competitor to AES systems in evaluating student essays automatically. Considering the potential limitations of AI in processing idioms, this study assessed the scoring performances of Generative AI models for essays with and without idioms by incorporating insights from Corpus Linguistics and Computational Linguistics. Two equal essay lists were created from 348 student essays taken from a corpus: one with multiple idioms present in each essay and another with no idioms in essays. Three Generative AI models (ChatGPT, Gemini, and Deepseek) were asked to score all essays in both lists three times, using the same rubric used by human raters in assigning essay scores. The results revealed excellent consistency for all models, but Gemini outperformed its competitors in interrater reliability with human raters. There was also no detectable bias for any demographic group in AI assessment. For essays with multiple idioms, Gemini followed a the most similar pattern to human raters. While the models in the study demonstrated potential for a hybrid approach, Gemini was the best candidate for the task due to its ability to handle figurative language and showed promise for handling essay-scoring tasks alone in the future.",
"comment": "",
"journal_ref": "Assessing Writing (2025), 66, 100981",
"doi": "10.1016/j.asw.2025.100981",
"primary": "cs.CL"
},
{
"id": "2510.18234v1",
"version_id": "http://arxiv.org/abs/2510.18234v1",
"title": "DeepSeek-OCR: Contexts Optical Compression",
"published": "2025-10-21T02:41:44Z",
"updated": "2025-10-21T02:41:44Z",
"authors": [
"Haoran Wei",
"Yaofeng Sun",
"Yukun Li"
],
"summary": "We present DeepSeek-OCR as an initial investigation into the feasibility of compressing long contexts via optical 2D mapping. DeepSeek-OCR consists of two components: DeepEncoder and DeepSeek3B-MoE-A570M as the decoder. Specifically, DeepEncoder serves as the core engine, designed to maintain low activations under high-resolution input while achieving high compression ratios to ensure an optimal and manageable number of vision tokens. Experiments show that when the number of text tokens is within 10 times that of vision tokens (i.e., a compression ratio < 10x), the model can achieve decoding (OCR) precision of 97%. Even at a compression ratio of 20x, the OCR accuracy still remains at about 60%. This shows considerable promise for research areas such as historical long-context compression and memory forgetting mechanisms in LLMs. Beyond this, DeepSeek-OCR also demonstrates high practical value. On OmniDocBench, it surpasses GOT-OCR2.0 (256 tokens/page) using only 100 vision tokens, and outperforms MinerU2.0 (6000+ tokens per page on average) while utilizing fewer than 800 vision tokens. In production, DeepSeek-OCR can generate training data for LLMs/VLMs at a scale of 200k+ pages per day (a single A100-40G). Codes and model weights are publicly accessible at http://github.com/deepseek-ai/DeepSeek-OCR.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2511.19333v1",
"version_id": "http://arxiv.org/abs/2511.19333v1",
"title": "Learning to Reason: Training LLMs with GPT-OSS or DeepSeek R1 Reasoning Traces",
"published": "2025-11-24T17:26:58Z",
"updated": "2025-11-24T17:26:58Z",
"authors": [
"Shaltiel Shmidman",
"Asher Fredman",
"Oleg Sudakov",
"Meriem Bendris"
],
"summary": "Test-time scaling, which leverages additional computation during inference to improve model accuracy, has enabled a new class of Large Language Models (LLMs) that are able to reason through complex problems by understanding the goal, turning this goal into a plan, working through intermediate steps, and checking their own work before answering . Frontier large language models with reasoning capabilities, such as DeepSeek-R1 and OpenAI's gpt-oss, follow the same procedure when solving complex problems by generating intermediate reasoning traces before giving the final answer. Today, these models are being increasingly used to generate reasoning traces that serve as high-quality supervised data for post-training of small and medium-sized language models to teach reasoning capabilities without requiring expensive human curation. In this work, we compare the performance of medium-sized LLMs on Math problems after post-training on two kinds of reasoning traces. We compare the impact of reasoning traces generated by DeepSeek-R1 and gpt-oss LLMs in terms of accuracy and inference efficiency.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2512.02556v1",
"version_id": "http://arxiv.org/abs/2512.02556v1",
"title": "DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models",
"published": "2025-12-02T09:25:14Z",
"updated": "2025-12-02T09:25:14Z",
"authors": [
" DeepSeek-AI",
"Aixin Liu",
"Aoxue Mei",
"Bangcai Lin",
"Bing Xue",
"Bingxuan Wang",
"Bingzheng Xu",
"Bochao Wu",
"Bowei Zhang",
"Chaofan Lin",
"Chen Dong",
"Chengda Lu",
"Chenggang Zhao",
"Chengqi Deng",
"Chenhao Xu",
"Chong Ruan",
"Damai Dai",
"Daya Guo",
"Dejian Yang",
"Deli Chen",
"Erhang Li",
"Fangqi Zhou",
"Fangyun Lin",
"Fucong Dai",
"Guangbo Hao",
"Guanting Chen",
"Guowei Li",
"H. Zhang",
"Hanwei Xu",
"Hao Li",
"Haofen Liang",
"Haoran Wei",
"Haowei Zhang",
"Haowen Luo",
"Haozhe Ji",
"Honghui Ding",
"Hongxuan Tang",
"Huanqi Cao",
"Huazuo Gao",
"Hui Qu",
"Hui Zeng",
"Jialiang Huang",
"Jiashi Li",
"Jiaxin Xu",
"Jiewen Hu",
"Jingchang Chen",
"Jingting Xiang",
"Jingyang Yuan",
"Jingyuan Cheng",
"Jinhua Zhu",
"Jun Ran",
"Junguang Jiang",
"Junjie Qiu",
"Junlong Li",
"Junxiao Song",
"Kai Dong",
"Kaige Gao",
"Kang Guan",
"Kexin Huang",
"Kexing Zhou",
"Kezhao Huang",
"Kuai Yu",
"Lean Wang",
"Lecong Zhang",
"Lei Wang",
"Liang Zhao",
"Liangsheng Yin",
"Lihua Guo",
"Lingxiao Luo",
"Linwang Ma",
"Litong Wang",
"Liyue Zhang",
"M. S. Di",
"M. Y Xu",
"Mingchuan Zhang",
"Minghua Zhang",
"Minghui Tang",
"Mingxu Zhou",
"Panpan Huang",
"Peixin Cong",
"Peiyi Wang",
"Qiancheng Wang",
"Qihao Zhu",
"Qingyang Li",
"Qinyu Chen",
"Qiushi Du",
"Ruiling Xu",
"Ruiqi Ge",
"Ruisong Zhang",
"Ruizhe Pan",
"Runji Wang",
"Runqiu Yin",
"Runxin Xu",
"Ruomeng Shen",
"Ruoyu Zhang",
"S. H. Liu",
"Shanghao Lu",
"Shangyan Zhou",
"Shanhuang Chen",
"Shaofei Cai",
"Shaoyuan Chen",
"Shengding Hu",
"Shengyu Liu",
"Shiqiang Hu",
"Shirong Ma",
"Shiyu Wang",
"Shuiping Yu",
"Shunfeng Zhou",
"Shuting Pan",
"Songyang Zhou",
"Tao Ni",
"Tao Yun",
"Tian Pei",
"Tian Ye",
"Tianyuan Yue",
"Wangding Zeng",
"Wen Liu",
"Wenfeng Liang",
"Wenjie Pang",
"Wenjing Luo",
"Wenjun Gao",
"Wentao Zhang",
"Xi Gao",
"Xiangwen Wang",
"Xiao Bi",
"Xiaodong Liu",
"Xiaohan Wang",
"Xiaokang Chen",
"Xiaokang Zhang",
"Xiaotao Nie",
"Xin Cheng",
"Xin Liu",
"Xin Xie",
"Xingchao Liu",
"Xingkai Yu",
"Xingyou Li",
"Xinyu Yang",
"Xinyuan Li",
"Xu Chen",
"Xuecheng Su",
"Xuehai Pan",
"Xuheng Lin",
"Xuwei Fu",
"Y. Q. Wang",
"Yang Zhang",
"Yanhong Xu",
"Yanru Ma",
"Yao Li",
"Yao Li",
"Yao Zhao",
"Yaofeng Sun",
"Yaohui Wang",
"Yi Qian",
"Yi Yu",
"Yichao Zhang",
"Yifan Ding",
"Yifan Shi",
"Yiliang Xiong",
"Ying He",
"Ying Zhou",
"Yinmin Zhong",
"Yishi Piao",
"Yisong Wang",
"Yixiao Chen",
"Yixuan Tan",
"Yixuan Wei",
"Yiyang Ma",
"Yiyuan Liu",
"Yonglun Yang",
"Yongqiang Guo",
"Yongtong Wu",
"Yu Wu",
"Yuan Cheng",
"Yuan Ou",
"Yuanfan Xu",
"Yuduan Wang",
"Yue Gong",
"Yuhan Wu",
"Yuheng Zou",
"Yukun Li",
"Yunfan Xiong",
"Yuxiang Luo",
"Yuxiang You",
"Yuxuan Liu",
"Yuyang Zhou",
"Z. F. Wu",
"Z. Z. Ren",
"Zehua Zhao",
"Zehui Ren",
"Zhangli Sha",
"Zhe Fu",
"Zhean Xu",
"Zhenda Xie",
"Zhengyan Zhang",
"Zhewen Hao",
"Zhibin Gou",
"Zhicheng Ma",
"Zhigang Yan",
"Zhihong Shao",
"Zhixian Huang",
"Zhiyu Wu",
"Zhuoshu Li",
"Zhuping Zhang",
"Zian Xu",
"Zihao Wang",
"Zihui Gu",
"Zijia Zhu",
"Zilin Li",
"Zipeng Zhang",
"Ziwei Xie",
"Ziyi Gao",
"Zizheng Pan",
"Zongqing Yao",
"Bei Feng",
"Hui Li",
"J. L. Cai",
"Jiaqi Ni",
"Lei Xu",
"Meng Li",
"Ning Tian",
"R. J. Chen",
"R. L. Jin",
"S. S. Li",
"Shuang Zhou",
"Tianyu Sun",
"X. Q. Li",
"Xiangyue Jin",
"Xiaojin Shen",
"Xiaosha Chen",
"Xinnan Song",
"Xinyi Zhou",
"Y. X. Zhu",
"Yanping Huang",
"Yaohui Li",
"Yi Zheng",
"Yuchen Zhu",
"Yunxian Ma",
"Zhen Huang",
"Zhipeng Xu",
"Zhongyu Zhang",
"Dongjie Ji",
"Jian Liang",
"Jianzhong Guo",
"Jin Chen",
"Leyi Xia",
"Miaojun Wang",
"Mingming Li",
"Peng Zhang",
"Ruyi Chen",
"Shangmian Sun",
"Shaoqing Wu",
"Shengfeng Ye",
"T. Wang",
"W. L. Xiao",
"Wei An",
"Xianzu Wang",
"Xiaowen Sun",
"Xiaoxiang Wang",
"Ying Tang",
"Yukun Zha",
"Zekai Zhang",
"Zhe Ju",
"Zhen Zhang",
"Zihua Qu"
],
"summary": "We introduce DeepSeek-V3.2, a model that harmonizes high computational efficiency with superior reasoning and agent performance. The key technical breakthroughs of DeepSeek-V3.2 are as follows: (1) DeepSeek Sparse Attention (DSA): We introduce DSA, an efficient attention mechanism that substantially reduces computational complexity while preserving model performance in long-context scenarios. (2) Scalable Reinforcement Learning Framework: By implementing a robust reinforcement learning protocol and scaling post-training compute, DeepSeek-V3.2 performs comparably to GPT-5. Notably, our high-compute variant, DeepSeek-V3.2-Speciale, surpasses GPT-5 and exhibits reasoning proficiency on par with Gemini-3.0-Pro, achieving gold-medal performance in both the 2025 International Mathematical Olympiad (IMO) and the International Olympiad in Informatics (IOI). (3) Large-Scale Agentic Task Synthesis Pipeline: To integrate reasoning into tool-use scenarios, we developed a novel synthesis pipeline that systematically generates training data at scale. This methodology facilitates scalable agentic post-training, yielding substantial improvements in generalization and instruction-following robustness within complex, interactive environments.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2512.09772v2",
"version_id": "http://arxiv.org/abs/2512.09772v2",
"title": "DeepSeek's WEIRD Behavior: The cultural alignment of Large Language Models and the effects of prompt language and cultural prompting",
"published": "2025-12-10T15:54:18Z",
"updated": "2025-12-12T17:25:30Z",
"authors": [
"James Luther",
"Donald Brown"
],
"summary": "Culture is a core component of human-to-human interaction and plays a vital role in how we perceive and interact with others. Advancements in the effectiveness of Large Language Models (LLMs) in generating human-sounding text have greatly increased the amount of human-to-computer interaction. As this field grows, the cultural alignment of these human-like agents becomes an important field of study. Our work uses Hofstede's VSM13 international surveys to understand the cultural alignment of the following models: DeepSeek-V3, V3.1, GPT-4, GPT-4.1, GPT-4o, and GPT-5. We use a combination of prompt language and cultural prompting, a strategy that uses a system prompt to shift a model's alignment to reflect a specific country, to align these LLMs with the United States and China. Our results show that DeepSeek-V3, V3.1, and OpenAI's GPT-5 exhibit a close alignment with the survey responses of the United States and do not achieve a strong or soft alignment with China, even when using cultural prompts or changing the prompt language. We also find that GPT-4 exhibits an alignment closer to China when prompted in English, but cultural prompting is effective in shifting this alignment closer to the United States. Other low-cost models, GPT-4o and GPT-4.1, respond to the prompt language used (i.e., English or Simplified Chinese) and cultural prompting strategies to create acceptable alignments with both the United States and China.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2512.10576v1",
"version_id": "http://arxiv.org/abs/2512.10576v1",
"title": "ESS: An Offload-Centric Latent-Cache Management Architecture for DeepSeek-V3.2-Exp",
"published": "2025-12-11T12:06:00Z",
"updated": "2025-12-11T12:06:00Z",
"authors": [
"Xinhang Chen",
"Chao Zhang",
"Jiahuan He",
"Wei Liu",
"Jianming Zhang",
"Wenlong Zhou",
"Xiao Li",
"Pai Zeng",
"Shiyong Li",
"Yuanpan Qian",
"Dong Li",
"Zhaogeng Li"
],
"summary": "DeepSeek-V3.2-Exp introduces a sparse attention mechanism that significantly reduces inference latency in long-context scenarios. Although the overall throughput has improved greatly, the Decode-stage of PD disaggregation remains to be a major bottleneck. This bottleneck primarily stems from the conflict between linear growth of Latent-Cache with sequence length and the limited GPU memory capacity, which constrains the feasible batch-size and thereby suppresses Decode-stage throughput. To address this challenge, we propose ESS (Extended Sparse Server), an offload-centric system design tailored for DeepSeek-V3.2-Exp. ESS selectively offloads Latent-Cache to CPU memory while preserving latency-critical components on GPU. By freeing up GPU memory, ESS effectively decoupling batch-size scaling from GPU memory constraints. This design significantly improves Decode-stage throughput, thereby reducing deployment costs in real-world settings. Our high-fidelity simulations show that ESS delivers 69.4\\% throughput improvement at 32K context length and up to 123\\% throughput improvement at 128K, demonstrating its effectiveness for large-context inference workloads. These results highlight ESS as a practical and scalable solution for long-context LLM serving.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.DC"
},
{
"id": "2512.20344v1",
"version_id": "http://arxiv.org/abs/2512.20344v1",
"title": "A DeepSeek-Powered AI System for Automated Chest Radiograph Interpretation in Clinical Practice",
"published": "2025-12-23T13:26:13Z",
"updated": "2025-12-23T13:26:13Z",
"authors": [
"Yaowei Bai",
"Ruiheng Zhang",
"Yu Lei",
"Xuhua Duan",
"Jingfeng Yao",
"Shuguang Ju",
"Chaoyang Wang",
"Wei Yao",
"Yiwan Guo",
"Guilin Zhang",
"Chao Wan",
"Qian Yuan",
"Lei Chen",
"Wenjuan Tang",
"Biqiang Zhu",
"Xinggang Wang",
"Tao Sun",
"Wei Zhou",
"Dacheng Tao",
"Yongchao Xu",
"Chuansheng Zheng",
"Huangxuan Zhao",
"Bo Du"
],
"summary": "A global shortage of radiologists has been exacerbated by the significant volume of chest X-ray workloads, particularly in primary care. Although multimodal large language models show promise, existing evaluations predominantly rely on automated metrics or retrospective analyses, lacking rigorous prospective clinical validation. Janus-Pro-CXR (1B), a chest X-ray interpretation system based on DeepSeek Janus-Pro model, was developed and rigorously validated through a multicenter prospective trial (NCT07117266). Our system outperforms state-of-the-art X-ray report generation models in automated report generation, surpassing even larger-scale models including ChatGPT 4o (200B parameters), while demonstrating reliable detection of six clinically critical radiographic findings. Retrospective evaluation confirms significantly higher report accuracy than Janus-Pro and ChatGPT 4o. In prospective clinical deployment, AI assistance significantly improved report quality scores, reduced interpretation time by 18.3% (P < 0.001), and was preferred by a majority of experts in 54.3% of cases. Through lightweight architecture and domain-specific optimization, Janus-Pro-CXR improves diagnostic reliability and workflow efficiency, particularly in resource-constrained settings. The model architecture and implementation framework will be open-sourced to facilitate the clinical translation of AI-assisted radiology solutions.",
"comment": "arXiv admin note: substantial text overlap with arXiv:2507.19493",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2601.03714v2",
"version_id": "http://arxiv.org/abs/2601.03714v2",
"title": "Visual Merit or Linguistic Crutch? A Close Look at DeepSeek-OCR",
"published": "2026-01-07T09:01:23Z",
"updated": "2026-01-08T08:37:59Z",
"authors": [
"Yunhao Liang",
"Ruixuan Ying",
"Bo Li",
"Hong Li",
"Kai Yan",
"Qingwen Li",
"Min Yang",
"Okamoto Satoshi",
"Zhe Cui",
"Shiwen Ni"
],
"summary": "DeepSeek-OCR utilizes an optical 2D mapping approach to achieve high-ratio vision-text compression, claiming to decode text tokens exceeding ten times the input visual tokens. While this suggests a promising solution for the LLM long-context bottleneck, we investigate a critical question: \"Visual merit or linguistic crutch - which drives DeepSeek-OCR's performance?\" By employing sentence-level and word-level semantic corruption, we isolate the model's intrinsic OCR capabilities from its language priors. Results demonstrate that without linguistic support, DeepSeek-OCR's performance plummets from approximately 90% to 20%. Comparative benchmarking against 13 baseline models reveals that traditional pipeline OCR methods exhibit significantly higher robustness to such semantic perturbations than end-to-end methods. Furthermore, we find that lower visual token counts correlate with increased reliance on priors, exacerbating hallucination risks. Context stress testing also reveals a total model collapse around 10,000 text tokens, suggesting that current optical compression techniques may paradoxically aggravate the long-context bottleneck. This study empirically defines DeepSeek-OCR's capability boundaries and offers essential insights for future optimizations of the vision-text compression paradigm. We release all data, results and scripts used in this study at https://github.com/dududuck00/DeepSeekOCR.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2601.11464v1",
"version_id": "http://arxiv.org/abs/2601.11464v1",
"title": "MHA2MLA-VLM: Enabling DeepSeek's Economical Multi-Head Latent Attention across Vision-Language Models",
"published": "2026-01-16T17:45:34Z",
"updated": "2026-01-16T17:45:34Z",
"authors": [
"Xiaoran Fan",
"Zhichao Sun",
"Tao Ji",
"Lixing Shen",
"Tao Gui"
],
"summary": "As vision-language models (VLMs) tackle increasingly complex and multimodal tasks, the rapid growth of Key-Value (KV) cache imposes significant memory and computational bottlenecks during inference. While Multi-Head Latent Attention (MLA) offers an effective means to compress the KV cache and accelerate inference, adapting existing VLMs to the MLA architecture without costly pretraining remains largely unexplored. In this work, we present MHA2MLA-VLM, a parameter-efficient and multimodal-aware framework for converting off-the-shelf VLMs to MLA. Our approach features two core techniques: (1) a modality-adaptive partial-RoPE strategy that supports both traditional and multimodal settings by selectively masking nonessential dimensions, and (2) a modality-decoupled low-rank approximation method that independently compresses the visual and textual KV spaces. Furthermore, we introduce parameter-efficient fine-tuning to minimize adaptation cost and demonstrate that minimizing output activation error, rather than parameter distance, substantially reduces performance loss. Extensive experiments on three representative VLMs show that MHA2MLA-VLM restores original model performance with minimal supervised data, significantly reduces KV cache footprint, and integrates seamlessly with KV quantization.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2601.20552v1",
"version_id": "http://arxiv.org/abs/2601.20552v1",
"title": "DeepSeek-OCR 2: Visual Causal Flow",
"published": "2026-01-28T12:46:07Z",
"updated": "2026-01-28T12:46:07Z",
"authors": [
"Haoran Wei",
"Yaofeng Sun",
"Yukun Li"
],
"summary": "We present DeepSeek-OCR 2 to investigate the feasibility of a novel encoder-DeepEncoder V2-capable of dynamically reordering visual tokens upon image semantics. Conventional vision-language models (VLMs) invariably process visual tokens in a rigid raster-scan order (top-left to bottom-right) with fixed positional encoding when fed into LLMs. However, this contradicts human visual perception, which follows flexible yet semantically coherent scanning patterns driven by inherent logical structures. Particularly for images with complex layouts, human vision exhibits causally-informed sequential processing. Inspired by this cognitive mechanism, DeepEncoder V2 is designed to endow the encoder with causal reasoning capabilities, enabling it to intelligently reorder visual tokens prior to LLM-based content interpretation. This work explores a novel paradigm: whether 2D image understanding can be effectively achieved through two-cascaded 1D causal reasoning structures, thereby offering a new architectural approach with the potential to achieve genuine 2D reasoning. Codes and model weights are publicly accessible at http://github.com/deepseek-ai/DeepSeek-OCR-2.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2602.14158v1",
"version_id": "http://arxiv.org/abs/2602.14158v1",
"title": "A Multi-Agent Framework for Medical AI: Leveraging Fine-Tuned GPT, LLaMA, and DeepSeek R1 for Evidence-Based and Bias-Aware Clinical Query Processing",
"published": "2026-02-15T14:17:27Z",
"updated": "2026-02-15T14:17:27Z",
"authors": [
"Naeimeh Nourmohammadi",
"Md Meem Hossain",
"The Anh Han",
"Safina Showkat Ara",
"Zia Ush Shamszaman"
],
"summary": "Large language models (LLMs) show promise for healthcare question answering, but clinical use is limited by weak verification, insufficient evidence grounding, and unreliable confidence signalling. We propose a multi-agent medical QA framework that combines complementary LLMs with evidence retrieval, uncertainty estimation, and bias checks to improve answer reliability. Our approach has two phases. First, we fine-tune three representative LLM families (GPT, LLaMA, and DeepSeek R1) on MedQuAD-derived medical QA data (20k+ question-answer pairs across multiple NIH domains) and benchmark generation quality. DeepSeek R1 achieves the strongest scores (ROUGE-1 0.536 +- 0.04; ROUGE-2 0.226 +-0.03; BLEU 0.098 -+ 0.018) and substantially outperforms the specialised biomedical baseline BioGPT in zero-shot evaluation. Second, we implement a modular multi-agent pipeline in which a Clinical Reasoning agent (fine-tuned LLaMA) produces structured explanations, an Evidence Retrieval agent queries PubMed to ground responses in recent literature, and a Refinement agent (DeepSeek R1) improves clarity and factual consistency; an optional human validation path is triggered for high-risk or high-uncertainty cases. Safety mechanisms include Monte Carlo dropout and perplexity-based uncertainty scoring, plus lexical and sentiment-based bias detection supported by LIME/SHAP-based analyses. In evaluation, the full system achieves 87% accuracy with relevance around 0.80, and evidence augmentation reduces uncertainty (perplexity 4.13) compared to base responses, with mean end-to-end latency of 36.5 seconds under the reported configuration. Overall, the results indicate that agent specialisation and verification layers can mitigate key single-model limitations and provide a practical, extensible design for evidence-based and bias-aware medical AI.",
"comment": "27 pages, 14 figures, 5 tables",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2603.26673v1",
"version_id": "http://arxiv.org/abs/2603.26673v1",
"title": "Can AI be a Teaching Partner? Evaluating ChatGPT, Gemini, and DeepSeek across Three Teaching Strategies",
"published": "2026-02-24T00:47:40Z",
"updated": "2026-02-24T00:47:40Z",
"authors": [
"Talita de Paula Cypriano de Souza",
"Shruti Mehta",
"Matheus Arataque Uema",
"Luciano Bernardes de Paula",
"Seiji Isotani"
],
"summary": "There are growing promises that Large Language Models (LLMs) can support students' learning by providing explanations, feedback, and guidance. However, despite their rapid adoption and widespread attention, there is still limited empirical evidence regarding the pedagogical skills of LLMs. This article presents a comparative study of popular LLMs, namely, ChatGPT, DeepSeek, and Gemini, acting as teaching agents. An evaluation protocol was developed, focusing on three pedagogical strategies: Examples, Explanations and Analogies, and the Socratic Method. Six human judges conducted the evaluations in the context of teaching the C programming language to beginners. The results indicate that LLM models exhibited similar interaction patterns in the pedagogical strategies of Examples and Explanations and Analogies. In contrast, for the Socratic Method, the models showed greater sensitivity to the pedagogical strategy and the initial prompt. Overall, ChatGPT and Gemini received higher scores, whereas DeepSeek obtained lower scores across the criteria, indicating differences in pedagogical performance across models.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CY"
},
{
"id": "2603.19253v1",
"version_id": "http://arxiv.org/abs/2603.19253v1",
"title": "A comprehensive study of LLM-based argument classification: from Llama through DeepSeek to GPT-5.2",
"published": "2026-02-25T11:17:24Z",
"updated": "2026-02-25T11:17:24Z",
"authors": [
"Marcin Pietroń",
"Filip Gampel",
"Jakub Gomułka",
"Andrzej Tomski",
"Rafał Olszowski"
],
"summary": "Argument mining (AM) is an interdisciplinary research field focused on the automatic identification and classification of argumentative components, such as claims and premises, and the relationships between them. Recent advances in large language models (LLMs) have significantly improved the performance of argument classification compared to traditional machine learning approaches. This study presents a comprehensive evaluation of several state-of-the-art LLMs, including GPT-5.2, Llama 4, and DeepSeek, on large publicly available argument classification corpora such as Args.me and UKP. The evaluation incorporates advanced prompting strategies, including Chain-of- Thought prompting, prompt rephrasing, voting, and certainty-based classification. Both quantitative performance metrics and qualitative error analysis are conducted to assess model behavior. The best-performing model in the study (GPT-5.2) achieves a classification accuracy of 78.0% (UKP) and 91.9% (Args.me). The use of prompt rephrasing, multi-prompt voting, and certainty estimation further improves classification performance and robustness. These techniques increase the accuracy and F1 metric of the models by typically a few percentage points (from 2% to 8%). However, qualitative analysis reveals systematic failure modes shared across models, including instabilities with respect to prompt formulation, difficulties in detecting implicit criticism, interpreting complex argument structures, and aligning arguments with specific claims. This work contributes the first comprehensive evaluation that combines quantitative benchmarking and qualitative error analysis on multiple argument mining datasets using advanced LLM prompting strategies.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2604.03476v2",
"version_id": "http://arxiv.org/abs/2604.03476v2",
"title": "Fine-tuning DeepSeek-OCR-2 for Molecular Structure Recognition",
"published": "2026-04-03T21:42:54Z",
"updated": "2026-04-21T00:23:18Z",
"authors": [
"Haocheng Tang",
"Xingyu Dang",
"Junmei Wang"
],
"summary": "Optical Chemical Structure Recognition (OCSR) is critical for converting 2D molecular diagrams from printed literature into machine-readable formats. While Vision-Language Models have shown promise in end-to-end OCR tasks, their direct application to OCSR remains challenging, and direct full-parameter supervised fine-tuning often fails. In this work, we adapt DeepSeek-OCR-2 for molecular optical recognition by formulating the task as image-conditioned SMILES generation. To overcome training instabilities, we propose a two-stage progressive supervised fine-tuning strategy: starting with parameter-efficient LoRA and transitioning to selective full-parameter fine-tuning with split learning rates. We train our model on a large-scale corpus combining synthetic renderings from PubChem and realistic patent images from USPTO-MOL to improve coverage and robustness. Our fine-tuned model, MolSeek-OCR, demonstrates competitive capabilities, achieving exact matching accuracies comparable to the best-performing image-to-sequence model. However, it remains inferior to state-of-the-art image-to-graph modelS. Furthermore, we explore reinforcement-style post-training and data-curation-based refinement, finding that they fail to improve the strict sequence-level fidelity required for exact SMILES matching.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2604.12548v1",
"version_id": "http://arxiv.org/abs/2604.12548v1",
"title": "DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection",
"published": "2026-04-14T10:20:15Z",
"updated": "2026-04-14T10:20:15Z",
"authors": [
"Junyu Ren",
"Xingjian Pan",
"Wensheng Gan",
"Philip S. Yu"
],
"summary": "Prompt injection has emerged as a critical security threat to large language models (LLMs), yet existing studies predominantly focus on single-dimensional attack strategies, such as semantic rewriting or character-level obfuscation, which fail to capture the combined effects of multi-space perturbations in realistic scenarios. In addition, systematic black-box robustness evaluations of recent Chinese LLMs, such as DeepSeek, remain limited. To address these gaps, we propose PromptFuzz-SC, a semantic-character dual-space mutation framework for evaluating LLM robustness against prompt injection. The framework integrates semantic transformations (e.g., paraphrasing and word-order perturbation) with character-level obfuscation (e.g., zero-width insertion and encoding-based mutation), forming a unified and extensible mutation operator library. A hybrid search strategy combining epsilon-greedy exploration and hill-climbing refinement is adopted to efficiently discover high-quality adversarial prompts. We further introduce a unified evaluation protocol based on three metrics: misuse success rate (MSR), Average Queries to Success (AQS), and Stealth. Experimental results on DeepSeek demonstrate that dual-space mutation achieves the strongest overall attack performance among the evaluated strategies, attaining the highest mean MSR (0.189), peak MSR (0.375), and mean Stealth. Compared with semantic-only and character-only mutation, it improves mean MSR by 12.5% and 5.6%, respectively. While not consistently minimizing query cost, the proposed method achieves competitive best-case efficiency and maintains strong imperceptibility, indicating a more favorable balance between attack effectiveness and concealment. These findings highlight the importance of composite mutation strategies for robust red-teaming of LLMs and provide practical insights for the design of multi-layer defense mechanisms.",
"comment": "Preprint",
"journal_ref": "",
"doi": "",
"primary": "cs.CR"
},
{
"id": "2606.19348v1",
"version_id": "http://arxiv.org/abs/2606.19348v1",
"title": "DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence",
"published": "2026-04-26T14:49:33Z",
"updated": "2026-04-26T14:49:33Z",
"authors": [
" DeepSeek-AI",
"Anyi Xu",
"Bangcai Lin",
"Bing Xue",
"Bingxuan Wang",
"Bingzheng Xu",
"Bochao Wu",
"Bowei Zhang",
"Chaofan Lin",
"Chen Dong",
"Chenchen Ling",
"Chengda Lu",
"Chenggang Zhao",
"Chengqi Deng",
"Chengyu Hou",
"Chenhao Xu",
"Chenze Shao",
"Chong Ruan",
"Conner Sun",
"Damai Dai",
"Daya Guo",
"Dejian Yang",
"Deli Chen",
"Donghao Li",
"Dongjie Ji",
"Erhang Li",
"Fang Wei",
"Fangyun Lin",
"Fangzhou Yuan",
"Feiyu Xia",
"Fucong Dai",
"Guangbo Hao",
"Guanting Chen",
"Guoai Cao",
"Guolai Meng",
"Guowei Li",
"Han Yu",
"Han Zhang",
"Hanwei Xu",
"Hao Li",
"Haofen Liang",
"Haoling Zhang",
"Haoming Luo",
"Haoran Wei",
"Haotian Yuan",
"Haowei Zhang",
"Haowen Luo",
"Haoyu Chen",
"Haozhe Ji",
"Hengqing Zhang",
"Honghui Ding",
"Hongxuan Tang",
"Huanqi Cao",
"Huazuo Gao",
"Hui Qu",
"Hui Zeng",
"J Yang",
"JQ Zhu",
"Jia Luo",
"Jia Song",
"Jia Yu",
"Jialiang Huang",
"Jialu Cai",
"Jian Liang",
"Jiangting Zhou",
"Jiasheng Ye",
"Jiashi Li",
"Jiaxin Xu",
"Jiewen Hu",
"Jieyu Yang",
"Jin Chen",
"Jin Yan",
"Jingchang Chen",
"Jingli Zhou",
"Jingting Xiang",
"Jingyang Yuan",
"Jingyuan Cheng",
"Jingzi Zhou",
"Jinhua Zhu",
"Jiping Yu",
"Joseph Sun",
"Jun Ran",
"Junguang Jiang",
"Junjie Qiu",
"Junlong Li",
"Junmin Zheng",
"Junxiao Song",
"Kai Dong",
"Kaige Gao",
"Kang Guan",
"Kexing Zhou",
"Kezhao Huang",
"Kuai Yu",
"Lean Wang",
"Lecong Zhang",
"Lei Wang",
"Leyi Xia",
"Li Zhang",
"Liang Zhao",
"Lihua Guo",
"Lingxiao Luo",
"Linwang Ma",
"Linyan Zhu",
"Litong Wang",
"Liyu Cai",
"Liyue Zhang",
"Longhao Chen",
"MS Di",
"MY Xu",
"Max Mei",
"Miaojun Wang",
"Mingchuan Zhang",
"Minghua Zhang",
"Minghui Tang",
"Mingming Li",
"Mingxu Zhou",
"Minmin Han",
"Ning Wang",
"Panpan Huang",
"Panpan Wang",
"Peixin Cong",
"Peiyi Wang",
"Peng Zhang",
"Qiancheng Wang",
"Qihao Zhu",
"Qingyang Li",
"Qinyu Chen",
"Qiushi Du",
"Qiwei Jiang",
"Rui Tian",
"Ruifan Xu",
"Ruijie Lu",
"Ruiling Xu",
"Ruiqi Ge",
"Ruisong Zhang",
"Ruizhe Pan",
"Runji Wang",
"Runqian Chen",
"Runqiu Yin",
"Runxin Xu",
"Ruomeng Shen",
"Ruoyu Zhang",
"Ruyi Chen",
"SH Liu",
"Shanghao Lu",
"Shangmian Sun",
"Shangyan Zhou",
"Shanhuang Chen",
"Shaofei Cai",
"Shaoheng Nie",
"Shaoqing Wu",
"Shaoyuan Chen",
"Shengding Hu",
"Shengyu Liu",
"Shiqiang Hu",
"Shirong Ma",
"Shiyu Wang",
"Shuiping Yu",
"Shunfeng Zhou",
"Shuting Pan",
"Shuying Yu",
"Songyang Zhou",
"Tao Ni",
"Tao Yun",
"Tian Jin",
"Tian Pei",
"Tian Ye",
"Tianle Lin",
"Tianran Ji",
"Tianyi Cui",
"Tianyuan Yue",
"Tingting Yu",
"Tun Wang",
"W Zhang",
"WL Xiao",
"Wangding Zeng",
"Wei An",
"Weilin Zhao",
"Wen Liu",
"Wenfeng Liang",
"Wenjie Pang",
"Wenjing Luo",
"Wenjing Yao",
"Wenjun Gao",
"Wenkai Yang",
"Wenlve Huang",
"Wenqing Hou",
"Wentao Zhang",
"Wenting Ma",
"Xi Gao",
"Xiang He",
"Xiangwen Wang",
"Xianzu Wang",
"Xiao Bi",
"Xiaodong Liu",
"Xiaohan Wang",
"Xiaokang Chen",
"Xiaokang Zhang",
"Xiaotao Nie",
"Xiaowen Sun",
"Xiaoxiang Wang",
"Xin Cheng",
"Xin Liu",
"Xin Xie",
"Xingchao Liu",
"Xingchen Liu",
"Xingkai Yu",
"Xingyou Li",
"Xinyu Yang",
"Xinyu Zhang",
"Xu Chen",
"Xuanyu Wang",
"Xuecheng Su",
"Xueyin Chen",
"Xuheng Lin",
"Xuwei Fu",
"YC Yan",
"YQ Wang",
"YW Ma",
"Yanfeng Luo",
"Yang Zhang",
"Yanhong Xu",
"Yanru Ma",
"Yanwen Huang",
"Yao Li",
"Yao Li",
"Yao Xu",
"Yao Zhao",
"Yaofeng Sun",
"Yaohui Wang",
"Yi Qian",
"Yi Shao",
"Yi Yu",
"Yichao Zhang",
"Yifan Ding",
"Yifan Shi",
"Yijia Wu",
"Yiliang Xiong",
"Yiling Ma",
"Ying He",
"Ying Tang",
"Ying Zhou",
"Yingjia Luo",
"Yinmin Zhong",
"Yishi Piao",
"Yisong Wang",
"Yixiang Zhang",
"Yixiao Chen",
"Yixuan Tan",
"Yixuan Wei",
"Yiyang Ma",
"Yiyuan Liu",
"Yonglun Yang",
"Yongqiang Guo",
"Yongtong Wu",
"Yu Wu",
"YuKun Li",
"Yuan Cheng",
"Yuan Ou",
"Yuanfan Xu",
"Yuanhao Li",
"Yuduan Wang",
"Yuehan Yang",
"Yuer Xu",
"Yuhan Wu",
"Yuhao Meng",
"Yuheng Zou",
"Yukun Zha",
"Yunfan Xiong",
"Yupeng Chen",
"Yuping Lin",
"Yuqian Cao",
"Yuqian Wang",
"Yushun Zhang",
"Yuting Yan",
"Yutong Lin",
"Yuxian Gu",
"Yuxiang Luo",
"Yuxiang You",
"Yuxuan Liu",
"Yuxuan Zhou",
"Yuyang Zhou",
"Yuzhen Huang",
"ZF Wu",
"Zehao Wang",
"Zehua Zhao",
"Zehui Ren",
"Zekai Zhang",
"Zhangli Sha",
"Zhe Fu",
"Zhe Ju",
"Zhean Xu",
"Zhenda Xie",
"Zhengyan Zhang",
"Zheren Gao",
"Zhewen Hao",
"Zhibin Gou",
"Zhicheng Ma",
"Zhigang Yan",
"Zhihong Shao",
"Zhixian Huang",
"Zhixuan Chen",
"Zhiyu Wu",
"Zhizhou Ren",
"Zhongyu Wu",
"Zhuoshu Li",
"Zhuping Zhang",
"Zian Xu",
"Zihao Wang",
"Zihua Qu",
"Zihui Gu",
"Zijia Zhu",
"Zilin Li",
"Zipeng Zhang",
"Ziwei Xie",
"Ziyi Gao",
"Ziyi Wan",
"Zizheng Pan",
"Zongqing Yao"
],
"summary": "We present a preview version of DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models -- DeepSeek-V4-Pro with 1.6T parameters (49B activated) and DeepSeek-V4-Flash with 284B parameters (13B activated) -- both supporting a context length of one million tokens. DeepSeek-V4 series incorporate several key upgrades in architecture and optimization: (1) a hybrid attention architecture that combines Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to improve long-context efficiency; (2) Manifold-Constrained Hyper-Connections (mHC) that enhance conventional residual connections; (3) and the Muon optimizer for faster convergence and greater training stability. We pre-train both models on more than 32T diverse and high-quality tokens, followed by a comprehensive post-training pipeline that unlocks and further enhances their capabilities. DeepSeek-V4-Pro-Max, the maximum reasoning effort mode of DeepSeek-V4-Pro, redefines the state-of-the-art for open models, outperforming its predecessors in core tasks. Meanwhile, DeepSeek-V4 series are highly efficient in long-context scenarios. In the one-million-token context setting, DeepSeek-V4-Pro requires only 27% of single-token inference FLOPs and 10% of KV cache compared with DeepSeek-V3.2. This enables us to routinely support one-million-token contexts, thereby making long-horizon tasks and further test-time scaling more feasible. The model checkpoints are available at https://huggingface.co/collections/deepseek-ai/deepseek-v4.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2605.00392v3",
"version_id": "http://arxiv.org/abs/2605.00392v3",
"title": "RTPrune: Reading-Twice Inspired Token Pruning for Efficient DeepSeek-OCR Inference",
"published": "2026-05-01T04:30:16Z",
"updated": "2026-05-21T12:40:15Z",
"authors": [
"Ben Wan",
"Yan Feng",
"Zihan Tang",
"Weizhe Huang",
"Yuting Zeng",
"Jia Wang",
"Tongxuan Liu"
],
"summary": "DeepSeek-OCR leverages visual-text compression to reduce long-text processing costs and accelerate inference, yet visual tokens remain prone to redundant textual and structural information. Moreover, current token pruning methods for conventional vision-language models (VLMs) fail to preserve textual fidelity due to improper compression mechanisms. By analyzing the decoding process of DeepSeek-OCR, we find that a distinct two-stage reading trajectory: the model initially prioritizes the majority of high-norm tokens, then subsequently redistributes its attention to the remaining ones. Motivated by this insight, we propose RTPrune, a two-stage token pruning method tailored for DeepSeek-OCR. In the first stage, we prioritize high-norm visual tokens that capture salient textual and structural information. In the second stage, the remaining tokens are paired and merged based on optimal transport theory to achieve efficient feature aggregation. We further introduce a dynamic pruning ratio that adapts to token similarity and textual density for OCR tasks, enabling a better efficiency-accuracy trade-off. Extensive experiments demonstrate state-of-the-art performance, as evidenced by 99.47% accuracy and 1.23$\\times$ faster prefill on OmniDocBench, achieved with 84.25% token retention when applied to DeepSeek-OCR-Large.",
"comment": "21 pages, accepted by ICML2026",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2605.08615v1",
"version_id": "http://arxiv.org/abs/2605.08615v1",
"title": "DSPE: An Energy-Efficient Edge Processor for DeepSeek Inference with MerkleTree-based Incremental Pruning, Multi-Stage Boothing Lookup and Dynamic Adaptive Posit Processing",
"published": "2026-05-09T02:18:59Z",
"updated": "2026-05-09T02:18:59Z",
"authors": [
"Yuhan Zhang",
"Zhou Wang",
"Zhou Shu",
"Jiuren Zhou",
"Yanqing Xu",
"Xiaonan Tang",
"Shushan Qiao",
"Tianchun Ye",
"Yang Liu",
"Anil A. Bharath",
"Emm Mic Drakakis"
],
"summary": "In recent years, DeepSeek has achieved strong inference performance but remains hard to deploy on energy-constrained edge devices. This paper presents the DeepSeek Processing Element (DSPE), an edge-oriented architecture that alleviates the model's heavy computational and energy demands. DSPE introduces three techniques: the MerkleTree-based Incremental Pruning Scheme (MIPS) for secure redundant-vector reduction, the Multi-Stage Boothing Lookup Method (MBLM) for bit-flip-aware approximate multiplication, and the Dynamic Adaptive Posit Processing Mechanism (DAPPM), which introduces a new DA-Posit format and its corresponding hardware multiplication architecture. Implemented in TSMC 28nm CMOS, DSPE achieves 109.4 TFLOPS/W energy efficiency compared with state-of-the-art designs and offers a scalable foundation for edge deployment.",
"comment": "Accepted by DAC 2026, Long Beach, CA, USA. 7 pages, 8 figures. Corresponding author: Zhou Wang. †These authors contributed equally to this work: Yuhan Zhang and Zhou Wang",
"journal_ref": "",
"doi": "",
"primary": "cs.AR"
},
{
"id": "2605.10234v1",
"version_id": "http://arxiv.org/abs/2605.10234v1",
"title": "Social Policy of Large Language Models: How GPT, Claude, DeepSeek and Grok Allocate Social Budgets in Spain and Germany",
"published": "2026-05-11T09:10:42Z",
"updated": "2026-05-11T09:10:42Z",
"authors": [
"Claudia Benavides Cantos",
"Eduardo C. Garrido-Merchán"
],
"summary": "We study how four widely used large language models, namely Claude, GPT-4o, DeepSeek and Grok, distribute a fixed national social budget across twelve macro-areas of public expenditure under two European national contexts, Spain and Germany. Each combination of model and country is queried six times under identical prompts and generation parameters, producing forty-eight independent allocations that are compared against approximate Organisation for Economic Co-operation and Development (OECD) reference budgets and against each other. We formalise five hypotheses regarding geopolitical bias, housing under-allocation, structural convergence, sensitivity to national context, and under-representation of politically sensitive categories. The differences between models are then validated through Kruskal-Wallis tests on each macro-area, with post-hoc Mann-Whitney U comparisons under Bonferroni correction, and complemented by an analysis of pairwise Pearson correlations and a lexical examination of the textual justifications produced by each model. The results show that all four models share a systematic implicit social policy that diverges from real European spending structures: pensions are under-allocated by a factor close to three, while housing and employment are over-allocated by factors of four and two respectively. The principal axis of differentiation between models is not geopolitical, since Claude and DeepSeek are the most correlated pair across both countries, but rather a contrast between concentration and dispersion of the budget. Only Claude exhibits substantive sensitivity to the national context. The conclusions delimit the conditions under which language models may responsibly support, but not replace, expert deliberation in public budgeting.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CY"
},
{
"id": "2606.01228v1",
"version_id": "http://arxiv.org/abs/2606.01228v1",
"title": "Institutional Trust and the Domestic AI Advantage: Evidence from DeepSeek and ChatGPT Users in China",
"published": "2026-05-31T13:25:46Z",
"updated": "2026-05-31T13:25:46Z",
"authors": [
"Jiashen Huang",
"Yu Jia",
"Xu Pan"
],
"summary": "Public trust in generative artificial intelligence exhibits increasingly divergent patterns across national contexts, yet prevailing research largely overlooks the macro-structural forces underlying this divergence. This study argues that trust in AI is not merely a technical response to performance but a product of institutional refraction. We propose an ``Institutional Prism'' framework to demonstrate how institutional trust shapes user trust in domestic (DeepSeek) and global (ChatGPT) large language models. Drawing on Cognitive-Affective Trust Theory, we distinguish between cognitive and affective dimensions of trust and analyze survey data from 405 Chinese users. The findings show that higher institutional trust is positively associated with stronger affective trust in domestic AI models and shifts cognitive evaluations in a more favorable direction. While under lower institutional trust, this domestic advantage weakens. These findings reveal that institutional trust has emerged as a core dimension of AI trust formation. By linking micro-level psychological judgments with macro-level governance, this research contributes a new perspective to human-machine communication.",
"comment": "48 pages",
"journal_ref": "",
"doi": "",
"primary": "cs.CY"
},
{
"id": "2606.07410v1",
"version_id": "http://arxiv.org/abs/2606.07410v1",
"title": "A Comprehensive Anatomy of Human and DeepSeek-R1 LLM Mathematical Reasoning",
"published": "2026-06-05T15:57:42Z",
"updated": "2026-06-05T15:57:42Z",
"authors": [
"Yuxiang Chen",
"Jun Wang"
],
"summary": "The emergence of \"Aha moments\" in large language models, particularly DeepSeek-R1-0120, has raised the question of whether these systems genuinely reason or merely imitate the appearance of reasoning. We conduct a comprehensive empirical comparison between model and human reasoning across all 30 problems from AIME 2025, exhaustively annotating 10,247 reasoning steps into five functional categories: Analysis, Inference, Branch, Backtrace, and Reflection. We find a clear structural difference. Human solutions maintain a compact alternation between analysis and deduction, whereas DeepSeek-R1 frequently revisits intermediate results, performs shallow and often unnecessary verification, and loops through local checks without meaningful logical progress. We describe this as topological mimicry: reproducing the surface form of reasoning without its functional role. Despite this, we identify two signals of genuine reasoning. First, successful traces exhibit stable use of branching and backtracking, while failed traces either underuse or overuse exploratory actions. Second, reflection is only effective when placed within deductive inference; reflections trapped in analysis loops focus on local numerical details while missing global logical errors. These findings suggest that current long-CoT models may be rewarded more for the appearance of reasoning than for genuine deductive progress. We discuss directions for improving evaluation and training, including measuring cross-trace stability, penalising \"spinning-wheel\" traces, encouraging deeper logical correction, and reallocating inference-time compute toward deduction and backtracking. Overall, reasoning quality depends not simply on how much reflection occurs, but on whether reflection appears consistently and at the appropriate logical scale.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.LG"
},
{
"id": "2606.09079v3",
"version_id": "http://arxiv.org/abs/2606.09079v3",
"title": "FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse Attention",
"published": "2026-06-08T06:25:54Z",
"updated": "2026-07-20T13:38:48Z",
"authors": [
"Yan Wang",
"Qifan Zhang",
"Jiachen Yu",
"Tian Liang",
"Dongyang Ma",
"Xiang Hu",
"Zibo Lin",
"Chunyang Li",
"Zhichao Wang",
"Miao Peng",
"Nuo Chen",
"Jia Li",
"Yujiu Yang",
"Haitao Mi",
"Dong Yu"
],
"summary": "Conventional LLMs keep the full KV cache loaded during decoding, causing a severe GPU memory bottleneck for ultra-long context serving. In this report, we propose \\textbf{Lookahead Sparse Attention (LSA)}, a novel inference paradigm powered by a Neural Memory Indexer built upon the DeepSeek-V4 architecture. Rather than passively attending to all historical tokens, LSA proactively predicts future context demands and preserves only the query-critical KV chunks in the GPU memory. Crucially, we instantiate this architecture via a \\textbf{backbone-free decoupled training} strategy. By formulating the indexer as a standard dual-encoder architecture, we train it independently using standard retrieval training frameworks without ever loading the massive backbone model into GPU memory. We demonstrate that this ``less is more'' paradigm significantly maximizes serving efficiency while acting as an effective attention denoiser in tasks that rely on long-term global memory. Across primary long-context evaluation suites (e.g., LongBench-v2, LongMemEval, and RULER), \\texttt{FM-DS-V4} compresses the average physical KV cache footprint down to merely 13.5\\% of the full-context baseline, while consistently preserving or slightly elevating downstream accuracy (+0.6\\% absolute margin on average). At 1M context, per-decode-token compute drops to 0.30$\\times$ of the baseline and GPU KV cache shrinks by 90\\% (3.73$\\to$0.37 GB), translating into \\textbf{2.8$\\times$ aggregate throughput and 2.7$\\times$ concurrency gains} in PD-disaggregated serving on 8$\\times$H20 GPUs.",
"comment": "Technical report. 11 pages. Code and model available at https://github.com/libertywing/FlashMemory-Deepseek-V4 and https://huggingface.co/libertywing/FlashMemory-Deepseek-V4",
"journal_ref": "",
"doi": "",
"primary": "cs.LG"
},
{
"id": "2606.10392v1",
"version_id": "http://arxiv.org/abs/2606.10392v1",
"title": "Instruction Finetuning DeepSeek-R1-8B Model Using LoRA and NEFTune",
"published": "2026-06-09T04:14:49Z",
"updated": "2026-06-09T04:14:49Z",
"authors": [
"Wu Yuerong",
"Mingni Luo"
],
"summary": "Financial named-entity recognition (NER) is essential for translating unstructured financial reports and news into structured knowledge graphs. However, general-purpose large language models (LLMs) often misclassify financial entities or ignore domain-specific patterns. This paper investigates the use of DeepSeek-R1-8B, a recent open-source large language model, combined with Low-Rank Adaptation (LoRA) and Noisy Embedding Fine-Tuning (NEFTune) for financial NER. Each annotated sentence in our corpus of 1693 samples is converted into an instruction-input-output triple. We insert lightweight LoRA matrices into the Transformer layers and apply NEFTune to improve generalisation by adding uniform noise to embedding vectors during training. Experiments show that the LoRA-adapted DeepSeek-R1-8B achieves a micro-F1 of 0.901 on seven entity types (Company, Date, Location, Money, Person, Product and Quantity), and adding NEFTune further boosts the micro-F1 to 0.912, outperforming Llama3-8B, Qwen3-8B, Baichuan2-7B, T5 and BERT-Base baselines.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2607.20145v3",
"version_id": "http://arxiv.org/abs/2607.20145v3",
"title": "SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD",
"published": "2026-07-22T13:49:17Z",
"updated": "2026-08-19T09:30:55Z",
"authors": [
"Dongfang Li",
"Xiaodong Luo",
"Ruoyu Sun",
"Xuhui Chen",
"Linyuan Qiu",
"Jian Meng",
"Zhengxuan Lu",
"Yiting Wang",
"Yucheng Xie",
"Tao Guo",
"Tianxiang Fang",
"Jing Li",
"Sihang Chen",
"Shihao Hong",
"Chang Liu",
"Weihua Dai",
"Zirong Zeng",
"Ziwei Zhu",
"Zhuohan Wang",
"Zhengjun Yue",
"Igor Vasilyev",
"Min Liu",
"Weijian Sun",
"Xin Chen",
"Yingmeng Gao",
"Jinhua Zhou",
"Taolue Chen",
"Chenwei Wu",
"Dong Zhang",
"Wenlong Jin",
"Jinmin Xiang",
"Barkova Maria",
"Ushakov Anton",
"Xianfei Jin",
"Tian Ding",
"Zhihang Lin",
"Qian Chen",
"Linxin Yang",
"Mingzhe Yang",
"Bingwei Zhang",
"Hongzhang Yang",
"Fangxue Zhang",
"Shijun Qin",
"Jie Yu",
"Cuihua Hu",
"Tolstykh Vasiliy",
"Nosov Ivan",
"Abdullin Amir",
"Zhicheng Zhou",
"Xin Zhang",
"Zhixiong Ning",
"Xutong Zhao",
"Junjie Huang",
"Jiajun Liu",
"Weiyan Kong",
"Zheng Zhang",
"Wenhan Luo",
"Lin Hu",
"Yangbo Guo",
"Li Zeng",
"Shihao Zhang",
"Baotian Hu",
"Min Zhang",
"Haizhou Li",
"Zhiquan Luo"
],
"summary": "Full-parameter post-training of trillion-parameter-scale MoE models introduces substantial system-level challenges for large-scale distributed training, including severe memory pressure, non-overlapped communication overhead, and inefficient kernel execution. While most large-scale LLM training systems are built around GPU-based clusters, this report presents an end-to-end optimization practice on the Ascend NPU SuperPOD. Using the DeepSeek-V4 model family as the target workload, we develop a hierarchical optimization framework spanning model-level parallelism, computation-communication orchestration, and low-level kernel execution. The resulting system achieves 34.22% Model FLOPs Utilization (MFU) with a 2.93x improvement over the open-source baseline recipe while maintaining training stability. Building on this optimized infrastructure, we further establish a CPT and SFT workflow for complex Operations Research (OR) tasks. We refer to the integrated framework as SLAI T-Rex. Using DeepSeek-V4-Flash, we develop OR-oriented CPT and SFT data pipelines that combine collected domain resources with solver-verified synthetic optimization documents. The resulting dataset contains 10K high-quality SFT samples spanning four task categories and three problem representations. The specialized model achieves the highest average zero-shot Pass@1 score among the evaluated models, reaching 71.81% and outperforming GPT-5.4-Mini and the base DeepSeek-V4-Flash model by 3.98 and 11.27 percentage points, respectively. Overall, this work demonstrates a full-stack pathway from efficient trillion-parameter model post-training on Ascend infra to domain-specialized Flash models for solver-grounded mathematical modeling, advancing frontier-model systems for complex reasoning.",
"comment": "73 pages, 22 figures, 20 tables",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
},
{
"id": "2608.16393v2",
"version_id": "http://arxiv.org/abs/2608.16393v2",
"title": "Security Assessment of DeepSeek Harness with A.I.G: Evaluating Resistance to Indirect Prompt Injection",
"published": "2026-08-17T10:43:07Z",
"updated": "2026-08-18T07:33:22Z",
"authors": [
"Zonghao Ying",
"Xiangfan Wu",
"Huiyu Wu",
"Xing Zheng",
"Huangsheng Cheng",
"Xiaorong Shi",
"Jing Guo"
],
"summary": "We assess indirect prompt injection in DeepSeek Harness (DSH), using AI-Infra-Guard (A.I.G) to construct tests, deliver controlled taint, execute DSH, collect traces, and judge outcomes. The study covers 14,560 controlled executions over 16 indirect-content channels, text and file carrier modes, 35 payload objectives, one unmodified baseline, and 12 attack methods. The experiment preserves DSH's agent loop, tool registry, model adapter, and session-event path; source tools and sensitive sinks are local fixtures, so attempted actions are recorded without external side effects. We evaluate each trace with a deterministic rule-based judge, \\JudgeR{} (RuleJudge), and a semantic LLM-based judge, \\JudgeL{} (LLMJudge). The strongest observed attack success rates are 17.0% under \\JudgeL{} for fake-completion attack in text mode, 25.5% under \\JudgeR{} for hidden Unicode in file mode, and 16.0% under \\JudgeR{} for the skills channel in file mode. \\JudgeL{} also assigns partial compliance more often than \\JudgeR{} (7.3% versus 2.0%). We relate these results to DSH's treatment of tool results, additional contexts, and tool-call policy hooks, then identify controls that should sit between untrusted content and sensitive actions. Our code is available at https://github.com/Tencent/AI-Infra-Guard/tree/main/Research/deepseek-harness-security-assessment .",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CR"
},
{
"id": "2608.30877v1",
"version_id": "http://arxiv.org/abs/2608.30877v1",
"title": "Deploying DeepSeek 175B Locally on a Single Consumer-Grade RTX 4060 Laptop with 32GB RAM for 200k-Scale Protein-Ligand Virtual Screening",
"published": "2026-08-31T14:35:56Z",
"updated": "2026-08-31T14:35:56Z",
"authors": [
"Rui Xiao",
"Yili Xu"
],
"summary": "Recent advances in large language models (LLMs) have demonstrated exceptional performance in protein-ligand interaction prediction, but state-of-the-art pipelines for large-scale virtual screening almost exclusively rely on high-end GPU clusters with hundreds of gigabytes of memory, creating prohibitive hardware barriers for small academic teams. In this work, we present a fully local low-resource framework that deploys the 175-billion-parameter DeepSeek 175B LLM on a single consumer-grade RTX 4060 laptop equipped with 32GB system RAM and 8GB VRAM, completing a full 200k-scale protein-ligand virtual screening workflow across 20 distinct protein targets. Our implementation achieves 100x throughput of an 8-card A100 cluster baseline under identical task configurations within 72 hours, with an average binding affinity prediction error of 0.88 kcal/mol across all targets, satisfying the 1.0 kcal/mol chemical accuracy requirement for preclinical drug discovery. Systematic runtime profiling reveals that heterogeneous memory management overhead accounts for 72% of total execution time, while accuracy loss introduced by model optimization contributes less than 10% to total prediction error. This work validates the engineering feasibility of running industrial-scale trillion-parameter LLM-driven biomedical computing tasks on consumer hardware, establishing a new low-barrier paradigm for AI-powered early stage drug discovery.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.LG"
},
{
"id": "2609.07008v1",
"version_id": "http://arxiv.org/abs/2609.07008v1",
"title": "RedKnot-MLA: Multi-Head Offline-Online Reuse for DeepSeek-V4 Long-Context Serving",
"published": "2026-09-07T03:59:33Z",
"updated": "2026-09-07T03:59:33Z",
"authors": [
"Yang Liu",
"Zhaokai Luo",
"Huayi Jin",
"Ruozhou He",
"Chenchen Hong",
"Mingxiao Ma",
"Biao Zhang",
"Zhiyong Wang",
"Boyu Wang",
"Guanjie Chen",
"Yifei Liu",
"Tao Xie",
"Junhao Hu"
],
"summary": "Multi-head latent attention (MLA) exposes many logical query heads through one packed latent KV stream. This representation is memory efficient, but it removes the physical per-head cache boundary assumed by conventional head-wise reuse. We present our system, a DeepSeek-V4 realization of RedKnot's head-aware reuse principle. Each immutable document is processed offline at canonical position zero; certified Local-head contributions are retained as MLA-Off. At serving time, query-side RoPE relocation restores the document's request position, a small Global-head set and protected Local token rows are recomputed as MLA-Online, and the two paths are merged before a single shared output projection. The packed MLA latent is never split. DeepSeek-V4-Flash uses 37 reusable layers and a 56/8 Local/Global partition, giving a 75.29% analytic logical head-row ceiling; the Pro-0813 profile uses 55 layers and 112/16 heads, giving 78.89%. Frozen Flash operating points show hot-artifact TTFT speedups of 2.02-3.84x. At 256K, the archived three-dataset study reports an aggregate F1 change of +3.24 percentage points, an EM change of +4.16 points, and a 78.7-79.5% analytic major-operator arithmetic saving, while one dataset decreases by 2.81 F1 points. A separate author-reported 256K hot-artifact QPS measurement is approximately 2.0x; because its raw concurrency trace is not included in this bundle, we mark it as preliminary rather than archived evidence. We describe the factorization, position repair, token-row closure, sparse-MoE support, TP8 integration, and the measurement boundaries needed to interpret these results.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.AI"
},
{
"id": "2609.15627v1",
"version_id": "http://arxiv.org/abs/2609.15627v1",
"title": "DeepSeek-V4-Flash on AMD gfx90a: Correctness Recovery and Inference Performance Engineering",
"published": "2026-09-14T14:18:39Z",
"updated": "2026-09-14T14:18:39Z",
"authors": [
"Siming Huang"
],
"summary": "We present the enablement, correctness recovery, and performance engineering of DeepSeek-V4-Flash inference on AMD Instinct MI250 GPUs using the gfx90a/CDNA2 architecture. The system integrates native safetensors loading, tensor and expert parallelism, FP4 routed mixture-of-experts computation, FP8 dense projections, sparse attention, HIP graph execution, and OpenAI-compatible serving within SGLang. An initially fast execution path was found to be numerically incorrect because of a routed-expert W2 layout mismatch. We identify the output permutation, repair the weight layout at load time, and establish fixed-token and hash-based correctness checks before further optimization. On the corrected path, decode performance is improved through packed FP4 weights, INT8 activation quantization, CDNA2 dot-product instructions, peer-read all-reduce, and topology-aware kernel geometry. Prefill is accelerated using CDNA2 MFMA kernels, improved packed-weight reuse, reduced sparse-attention overhead, larger chunks, and retuned expert sorting. On four MI250 GCDs, TP4/EP1 native autoregressive decode reaches approximately 74.5 tok/s, while a 4,604-token prompt reaches 2.061-2.062 s TTFT, or approximately 2,234 input tok/s. The results show that efficient DeepSeek-V4-Flash inference on CDNA2 is limited not only by memory bandwidth, but also by FP4 execution-format mismatch, low-M utilization, and per-layer synchronization costs.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.DC"
},
{
"id": "2609.18038v1",
"version_id": "http://arxiv.org/abs/2609.18038v1",
"title": "CoAtNet-DeepMoE: A Convolution-Attention Hybrid with DeepSeek Mixture-of-Experts for Parameter-Efficient Tomato Disease Classification",
"published": "2026-09-16T02:39:18Z",
"updated": "2026-09-16T02:39:18Z",
"authors": [
"Md Nadim Mahamood",
"Md Arif Shahriar",
"Md Shafi Ud Doula",
"Kamrul Hasan"
],
"summary": "The world population is growing rapidly, and technology is improving in parallel. Meeting the huge demand for food for these 7 billion people not only depends on increasing food production but also on reducing food loss. Crop losses due to disease affect both the food supply and the financial and economic stability of a country. Tomatoes are among the top food-producing crops globally, and a significant portion of this production is lost due to disease. People have used Machine Learning techniques for feature extraction and early diagnosis of tomato diseases, and nowadays, Deep Learning-based models are widely used for disease recognition. However, most existing models are highly parameter-intensive, which increases the time required for training and inference. As a result, while lightweight models are more suitable for user-friendly applications, they often show a reduction in performance. To balance performance and model size, we propose CoAtNet-DeepMoE, a Convolution-Attention hybrid architecture for rich feature extraction, further enhanced with a DeepSeek Mixture of Experts to substantially reduce the number of parameters without sacrificing accuracy. We evaluate our model on both balanced and imbalanced datasets from Kaggle and PlantVillage, demonstrating robustness and achieving 99.80% accuracy, 99.80% precision, 99.80% recall, and 99.80% F1-score on Kaggle, and 99.83% accuracy, 99.85% precision, 99.76% recall, and 99.80% F1-score on PlantVillage, representing state-of-the-art performance with only 2.47M parameters. The source code will be available at https://github.com/nadimbrur/CoAt-MoE.",
"comment": "16 pages, 8 tables, 6 figures",
"journal_ref": "",
"doi": "",
"primary": "cs.CV"
},
{
"id": "2609.19969v1",
"version_id": "http://arxiv.org/abs/2609.19969v1",
"title": "DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression",
"published": "2026-09-17T09:43:10Z",
"updated": "2026-09-17T09:43:10Z",
"authors": [
" DeepSeek-AI",
" :",
"Anyi Xu",
"B. Li",
"Bangcai Lin",
"Bing Xue",
"BingCheng Xian",
"Bingzheng Xu",
"Bochao Wu",
"Bowei Zhang",
"Boyi Deng",
"C. C. Yu",
"Chao Jin",
"Chaofan Lin",
"Chen Dong",
"Chenbing Wang",
"Chenfan Feng",
"Chengda Lu",
"Chenggang Zhao",
"Chengqi Deng",
"Chengyuan Zhang",
"Chenhao Xu",
"Chenqi Zhao",
"Chenze Shao",
"Chuhao Wang",
"Chuqi Zhang",
"Damai Dai",
"Dejian Yang",
"Deli Chen",
"Di Huang",
"Di Wu",
"Donghao Li",
"Erhang Li",
"Eric Fu",
"F. Zhou",
"Fangwei Zhou",
"Fangyun Lin",
"Fangzhou Yuan",
"Feiyu Xia",
"Fucong Dai",
"Guangbo Hao",
"Guanglin Li",
"Guanting Chen",
"Guoai Cao",
"Guofan Fan",
"Guolai Meng",
"Guowei Li",
"Haichuan Zhang",
"Haiyang Ma",
"Haiyang Shen",
"Han Li",
"Han Yu",
"Han Zhang",
"Hangyuan Deng",
"Hanwei Xu",
"Hanxiang Xu",
"Hanxun Zhong",
"Hao Guo",
"Hao Jiang",
"Hao Li",
"Hao Qin",
"Haodong Wen",
"Haofen Liang",
"Haofeng Huang",
"Haohua Liu",
"Haoling Zhang",
"Haoming Luo",
"Haoran Yang",
"Haotian Xu",
"Haotian Yuan",
"Haoting Huang",
"Haowen Luo",
"Haoyang Cai",
"Haoyu Chen",
"Haozhe Ji",
"Hengran Zhang",
"Hengrui Wang",
"Hengxu Wu",
"Honghui Ding",
"Hongxuan Tang",
"Huadong Wang",
"Huanqi Cao",
"Huazuo Gao",
"Hui Qu",
"Hui Zeng",
"J. Yang",
"J. H. Jin",
"J. H. Zhang",
"J. X. Zou",
"Jia Yu",
"Jiahui Zhou",
"Jiajun Chen",
"Jialiang Huang",
"Jialin Zhao",
"Jiamin Tang",
"Jian Zhou",
"Jianan Tong",
"Jianwen Li",
"Jiaqi Zhu",
"Jiarui Wang",
"Jiasheng Ye",
"Jiashi Li",
"Jiaxin Xu",
"Jiaying Ding",
"Jibai Lu",
"Jiewen Hu",
"Jin Yan",
"Jincheng Zhai",
"Jingchang Chen",
"Jingcheng Hu",
"Jingli Zhou",
"Jingsheng Xu",
"Jingting Xiang",
"Jingyan Yun",
"Jingyang Yuan",
"Jingyuan Cheng",
"Jinhua Zhu",
"Jinpeng Wang",
"Jinyi Chen",
"Jinyi Hu",
"Jiping Yu",
"Jueliang Guo",
"Junbo Pei",
"Junbo Sun",
"Junguang Jiang",
"Junjie Qiu",
"Junkang Zhou",
"Junqi Liu",
"Junren Li",
"Junxian Li",
"Junxiao Song",
"Junyi Guo",
"Kai Dong",
"Kaifeng Chen",
"Kaige Gao",
"Kang Guan",
"Kangdong Yuan",
"Ke Hong",
"Ke Xu",
"Kefan Zhao",
"Kexin Ji",
"Kexin Zhang",
"Kexing Zhou",
"Kuai Yu",
"Lan Zhang",
"Lean Wang",
"Lecong Zhang",
"Lei Wang",
"Letian Gao",
"Liang Zhao",
"Liansheng Xu",
"Lihua Guo",
"Lingxiao Luo",
"Lingyue Fu",
"Litao Deng",
"Litong Wang",
"Liyue Zhang",
"Longhao Chen",
"Lu Chen",
"Luotian Huang",
"Luyao Ma",
"Luyao Wang",
"M. S. Di",
"Max Mei",
"Menghao Ye",
"Miao Cui",
"Mingchuan Zhang",
"Minghua Zhang",
"Minghui Tang",
"Mingjing Zhang",
"Mingqi Wei",
"Mingshu Chen",
"Mingxing Liu",
"Mingxu Zhou",
"Mingyu Xu",
"Mingyu Yang",
"Mingze Wang",
"Muyang Chen",
"Ni Shentu",
"Ning Wang",
"Niufang Ning",
"Panpan Huang",
"Peixin Cong",
"Peiyi Wang",
"Peiyuan Xin",
"Pengfei Ren",
"Pengfei Yan",
"Pengle Zhang",
"Qi Kang",
"Qi Tang",
"Qiancheng Wang",
"Qiang Li",
"Qihao Zhu",
"Qingyang Li",
"Qinyu Chen",
"Qiushi Du",
"Qizhou Guo",
"Rongxian Xu",
"Rui Ding",
"Rui Hu",
"Rui Tian",
"Rui Yu",
"Ruidong Zhu",
"Ruifan Xu",
"Ruihan Yang",
"Ruihang Xia",
"Ruijie Lu",
"Ruilin Geng",
"Ruipeng Hong",
"Ruiqi Ge",
"Ruisong Zhang",
"Ruize Sun",
"Ruizhe Pan",
"Runji Wang",
"Runqian Chen",
"Runxin Xu",
"Ruohong Tian",
"Ruomeng Shen",
"Ruoyu Zhang",
"Ryan X.",
"S. H. Liu",
"Shanghao Lu",
"Shangyan Zhou",
"Shanhuang Chen",
"Shaofei Cai",
"Shaoheng Nie",
"Shaoyuan Chen",
"Shengding Hu",
"Shengkai Lin",
"Shengwen Ran",
"Shengyu Liu",
"Shengyuan Jia",
"Shi Bai",
"Shi Feng",
"Shicheng Xu",
"Shichun Liu",
"Shiqiang Hu",
"Shirong Ma",
"Shiyu Wang",
"Shiyuan Feng",
"Shufan Gong",
"Shuhan Lin",
"Shuiping Yu",
"Shunfeng Zhou",
"Shuo Yang",
"Shuomeng Wang",
"Shuting Guo",
"Shuting Pan",
"Shuying Yu",
"Sinuo Cao",
"Siyi Lin",
"Sizhe Chen",
"Songyang Chen",
"Songyang Zhou",
"Tao Ni",
"Tao Yun",
"Tian Jin",
"Tian Pei",
"Tian Ye",
"Tianle Lin",
"Tianran Ji",
"Tianyi Cui",
"Tianyuan Yue",
"Tingting Yu",
"Tongrui Xiong",
"Wangding Zeng",
"Wei Liu",
"Wei Zhang",
"Weibin Xu",
"Weihao Zeng",
"Weilin Zhao",
"Wen Liu",
"Wenfeng Liang",
"Wenjie Pang",
"Wenjing Luo",
"Wenjing Yao",
"Wenjun Gao",
"Wenkai Shao",
"Wenkai Yang",
"Wenli Zhang",
"Wenlu Wang",
"Wenlve Huang",
"Wenqian Yan",
"Wentao Zhang",
"Xi Gao",
"Xiang He",
"Xiang Li",
"Xiangli Li",
"Xiangwen Wang",
"Xiangying Zhang",
"Xiankui Wei",
"Xiao Bi",
"Xiaodong Liu",
"Xiaohan Wang",
"Xiaojian Qu",
"Xiaokang Chen",
"Xiaokang Zhang",
"Xiaotao Nie",
"Xiaoyao Zou",
"Xiaoyuan Li",
"Xicheng Guo",
"Xieting Chu",
"Xin Cheng",
"Xin Liu",
"Xin Xie",
"Xinbo Xu",
"Xingchao Liu",
"Xingchen Liu",
"Xingkai Yu",
"Xingyou Li",
"Xintong Yao",
"Xinyang Chen",
"Xinyong Jiang",
"Xinyu Yang",
"Xinyu Yang",
"Xu Chen",
"Xuanyu Wang",
"Xubei Zhong",
"Xuecheng Su",
"Xuejie Liu",
"Xuheng Lin",
"Xujie Fan",
"Xuncheng Zhao",
"Xuwei Fu",
"Y. C. Yan",
"Y. H. Jiang",
"Y. T. Wu",
"Y. W. M.",
"Y. Z. Wang",
"Yafei Gao",
"Yang Yang",
"Yang Zhang",
"Yanru Ma",
"Yanwen Huang",
"Yao Li",
"Yao Li",
"Yao Meng",
"Yao Zhao",
"Yaofeng Sun",
"Yaohui Wang",
"Yaoyang Ye",
"Yehang Yin",
"Yexinrui Wu",
"Yi Qian",
"Yi Tao",
"Yi Yu",
"Yichao Zhang",
"Yichen Jiang",
"Yicheng Wang",
"Yifan Ding",
"Yifan Shi",
"Yifeng Peng",
"Yifeng Zhai",
"Yijia Wu",
"Yiliang Xiong",
"Yilun Wang",
"Ying He",
"Ying Zhou",
"Yingjia Luo",
"Yinmin Zhong",
"Yiping Wang",
"Yisong Wang",
"Yixiang Zhang",
"Yixiao Chen",
"Yixuan Tan",
"Yixuan Wei",
"Yiyang Ma",
"Yiyao Yang",
"Yiyuan Liu",
"Yizai Cai",
"Yizhen Wei",
"Yizhi Wang",
"Yonglun Yang",
"Yongqi Zhuo",
"Yongqiang Guo",
"Yongtong Wu",
"Yu Wu",
"Yu Zhang",
"Yuan Bian",
"Yuan Cheng",
"Yuan Ou",
"Yuan Sun",
"Yuanfan Xu",
"Yuanhang Sun",
"Yuanhao Li",
"Yuchen Liu",
"Yuchen Yao",
"Yudong Han",
"Yuduan Wang",
"Yuhan Wu",
"Yuhao Meng",
"Yuheng Zou",
"YuKun Li",
"Yunchuan Wang",
"Yunfan Xiao",
"Yunfan Xiong",
"Yupeng Chen",
"Yuqian Cao",
"Yuqian Wang",
"Yuqing Chen",
"Yushun Zhang",
"Yutong Lin",
"Yuwei Xiao",
"Yuxian Gu",
"Yuxiang Chen",
"Yuxiang Huang",
"Yuxiang Luo",
"Yuxiang You",
"Yuxin Chen",
"Yuxin Xiang",
"Yuxuan Liu",
"Yuxuan Zhou",
"Yuyang Zhou",
"Yuzhe Guo",
"Yuzhen Huang",
"Yuzhuo Bai",
"Z. Y. Z.",
"Zanlin Ni",
"Zehao Wang",
"Zehua Zhao",
"Zehui Ren",
"Zejun Zhao",
"Zhangli Sha",
"Zhanying Wang",
"Zhaochen Zhang",
"Zhaoshuai Du",
"Zhe Fu",
"Zhean Xu",
"Zhenda Xie",
"Zheng Liu",
"Zhengyan Zhang",
"Zhenhua Dong",
"Zhewen Hao",
"Zhibang Wang",
"Zhibin Gou",
"Zhicheng Ma",
"Zhihao Li",
"Zhihong Shao",
"Zhihuan Huang",
"Zhijie Li",
"Zhirui Lu",
"Zhixian Huang",
"Zhixuan Chen",
"Zhixuan Chen",
"Zhixuan Pan",
"Zhiyu Wu",
"Zhizhou Ren",
"Zhu He",
"Zhuoshu Li",
"Zhuping Zhang",
"Zian Xu",
"Zihao Wang",
"Zihui Gu",
"Zijia Zhu",
"Zili Zhang",
"Zilin Li",
"Zilong Hou",
"Zilong Lyu",
"Ziqiao Wang",
"Ziwei Xie",
"Ziya Zhang",
"Ziyi Gao",
"Zizheng Pan",
"Zonglin Li",
"Zongqing Yao",
"Zui Chen",
"Zuofan Wu",
"Chenchen Ling",
"Chengyu Hou",
"Chong Chen",
"D. Li",
"Di Qi",
"Dongjie Ji",
"Fang Wei",
"Fanyi Xia",
"Fei Xie",
"Feiyi Tan",
"Hailong Guo",
"Haiyan Zhai",
"Hui Zhou",
"Huihui Tan",
"Huijie Li",
"Jia Luo",
"Jia Song",
"Jialu Cai",
"Jian Liang",
"Jiangting Zhou",
"Jiaqi Gao",
"Jiayi Shao",
"Jie Chen",
"Jieyu Yang",
"Jin Chen",
"Jingde Zhang",
"Jingzi Zhou",
"Jinqian Wang",
"Jinyang Liu",
"JinZhao Sun",
"Junhua Ling",
"Junmin Zheng",
"Kaicheng Yang",
"Ke Xu",
"Le Su",
"Leyi Xia",
"Liangfeng Ding",
"Lin Zhuo",
"Linwang Ma",
"Linyan Zhu",
"Liyu Cai",
"Luqi Yao",
"M. K. Zhang",
"Meng Li",
"Miao Lin",
"Miaojun Wang",
"Min Zhang",
"Mingming Li",
"Mingming Wang",
"Mingze Yin",
"Minmin Han",
"Nan Cao",
"Ning Wang",
"Ningxin Ma",
"Panpan Wang",
"Peihan Lin",
"Peng Sun",
"Peng Zhang",
"Qian Ying",
"Qiang Xiang",
"Qiao Wang",
"Qingmiao Mao",
"Qiwei Jiang",
"Rongli Jin",
"Ruyi Chen",
"Sha Tao",
"Shangmian Sun",
"Shaoqing Wu",
"Shichao Zou",
"Si Lei",
"Tianyang Zhang",
"Tianyu Sun",
"Tingting Yin",
"W. L. Xiao",
"Wei An",
"Wei Li",
"Wei Wang",
"Weiwei Lin",
"Wenqing Hou",
"X. Lin",
"Xiangfei Meng",
"Xianzhu Huang",
"Xiao Peng",
"Xiaoqian Li",
"Xiaoting Zhang",
"Xiaowen Sun",
"Xiaoxiang Wang",
"Xiaoyu Ye",
"Xinrou Zhang",
"Xinyu Zhang",
"Xue Cao",
"Xueyin Chen",
"Yanan Zhou",
"Yanhong Xu",
"Yao Xia",
"Yao Xu",
"Yi Shao",
"Yihong Zhang",
"Yiling Ma",
"Ying Tang",
"Yining Lou",
"Yiru Chen",
"Yishi Piao",
"Yixuan Chen",
"Yong Xiong",
"Yuchen Xuan",
"Yuehan Yang",
"Yuer Xu",
"Yukun Zha",
"Yunxian Ma",
"Yuping Lin",
"Yuting Yan",
"Yutong Xie",
"Yuwen Sheng",
"Yuxuan Zhu",
"Zekai Zhang",
"Zhe Ju",
"Zhenzhen Lin",
"Zheren Gao",
"Zheyang Sun",
"Zhigang Yan",
"Zhongyu Wu",
"Zi Wang",
"Zihua Qu",
"Ziling Yan",
"Ziyi Wan"
],
"summary": "The widespread adoption of long-horizon agents has made model workloads increasingly input-heavy. Although prior work has substantially reduced the cost of long-context computation, prefill remains computationally expensive, and large KV caches continue to strain HBM and SSD capacity and data-transfer bandwidth. Together, these compute, storage, and bandwidth demands constitute the primary bottleneck to further lowering deployment costs. To address this challenge, we introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. With its Causal Encoder-Decoder (CED) architecture, the model activates 16B parameters per token during decode but only 8B parameters during prefill, substantially improving cost efficiency for agentic workloads. To push the limits of KV cache compression, DeepSeek-V4.1-Flash combines cross-layer KV cache reuse in Compressed Sparse Attention 2 (CSA2) with FP4 KV caching. These designs reduce its global KV cache footprint (always in HBM) to 890 bytes per token, roughly 1/4 of the corresponding footprint of DeepSeek-V4-Flash. Further, through a dedicated deployment optimization known as SWA Bounded Replay, DeepSeek-V4.1-Flash reduces its persistent KV cache footprint (always on SSD or in host memory) to roughly 1/8 of that of DeepSeek-V4-Flash. Despite its much smaller KV cache footprint, the model delivers substantially better performance than the baseline. In addition, we streamline the DeepSeek-V4 architecture and introduce several efficient architectural extensions. We pretrain DeepSeek-V4.1-Flash on a multimodal corpus comprising 45T tokens and conduct comprehensive post-training, yielding strong performance across diverse text-based and multimodal agentic scenarios. Model checkpoints are available at https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash.",
"comment": "",
"journal_ref": "",
"doi": "",
"primary": "cs.CL"
}
]
Reference in New Issue View Git Blame Copy Permalink