大模型基准论文趋势 2023–2026

主要结论

• 大模型基准论文从 2023 年的 2 篇增长到 2026 年的 307 篇，三年间扩大到约 154 倍。
• NeurIPS 是发大模型基准论文最多的会议，2023–2026 年共 256 篇，占四个会议总量的 47%。
• 交叉方向里涨得最快的是AI 智能体：同时属于两个主题的论文从 2023 年的 1 篇增至 2026 年的 116 篇。
• 在「评测与基准」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的模型即裁判是 27 倍，大模型基准是 154 倍。

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

• NeurIPS 2026A Free Lunch in LLM Compression: Revisiting Retraining after Pruning
• ICML 2026Controlled LLM Training on Spectral Sphere
• ICLR 2026Hey, That's My Model! Introducing Chain & Hash, An LLM Fingerprinting Technique
• CVPR 20263DrawAgent Teaching LLM to Draw in 3D with Early Contrastive
• NeurIPS 2025ACCO: Accumulate While You Communicate for Communication-Overlapped Sharded LLM Training
• ICML 2025Critical Tokens Matter: Token-Level Contrastive Estimation Enhances LLM’s Reasoning Capability
• ICLR 2025Catastrophic Failure of LLM Unlearning via Quantization
• CVPR 2025Period-LLM: Extending the Periodic Capability of Multimodal Large Language Model
• NeurIPS 2024Large Language Models' Expert-level Global History Knowledge Benchmark (HiST-LLM)
• ICML 2024Characterizing ResNet's Universal Approximation Capability
• ICLR 2024Faithful Explanations of Black-box NLP Models Using LLM-generated Counterfactuals
• CVPR 2024Koala: Key Frame-Conditioned Long Video-LLM
• NeurIPS 2023Quantizable Transformers: Removing Outliers by Helping Attention Heads Do Nothing
• NeurIPS 2026Accelerating LLM Pre-Training through Flat-Direction Dynamics Enhancement
• ICML 2026LLM Priors for ERM over Programs
• ICLR 2026How Learning Rate Decay Wastes Your Best Data in Curriculum-Based LLM Pretraining
• CVPR 20263D-VCD Hallucination Mitigation in 3D-LLM Embodied Agents through Visual Contrastive
• NeurIPS 2025BrainEC-LLM: Brain Effective Connectivity Estimation by Multiscale Mixing LLM
• ICML 2025Position: Medical Large Language Model Benchmarks Should Prioritize Construct Validity
• ICLR 2025Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates

全部大模型基准论文

共 307 篇，第 2 / 7 页

51. ICLR 2026Log Probability Tracking of LLM APIs
52. ICLR 2026LS-Merge: Merging Language Models in Latent Space
53. ICLR 2026LycheeDecode: Accelerating Long-Context LLM Inference via Hybrid-Head Sparse Decoding
54. ICLR 2026Making Slow Thinking Faster: Compressing LLM Chain-of-Thought via Step Entropy
55. ICLR 2026MCP Security Bench (MSB): Benchmarking Attacks Against Model Context Protocol in LLM Agents
56. ICLR 2026MCP-Bench: Benchmarking Tool-Using LLM Agents with Complex Real-World Tasks via MCP Servers
57. ICLR 2026Metis: Training LLMs with FP4 Quantization
58. ICLR 2026Mix-Ecom: Towards Mixed-Type E-Commerce Dialogues with Complex Domain Rules
59. ICLR 2026NetArena: Dynamically Generated LLM Benchmarks for Network Applications
60. ICLR 2026NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents
61. ICLR 2026No Prompt Left Behind: Exploiting Zero-Variance Prompts in LLM Reinforcement Learning via Entropy-Guided Advantage Shaping
62. ICLR 2026On Entropy Control in LLM-RL Algorithms
63. ICLR 2026Operationalizing Data Minimization for Privacy-Preserving LLM Prompting
64. ICLR 2026PersonaX: Multimodal Datasets with LLM-Inferred Behavior Traits
65. ICLR 2026PILOT-Bench: Probabilistic Interaction for LLM Operations in Tool-driven Scenarios
66. ICLR 2026Preference Leakage: A Contamination Problem in LLM-as-a-judge
67. ICLR 2026Prompt Curriculum Learning for Efficient LLM Post-Training
68. ICLR 2026Randomization Boosts KV Caching, Learning Balances Query Load: A Joint Perspective
69. ICLR 2026Randomized Antipodal Search Done Right for Data Pareto Improvement of LLM Unlearning
70. ICLR 2026Rating Quality of Diverse Time Series Data by Meta-learning from LLM Judgment
71. ICLR 2026RCPU: Rotation-Constrained Error Compensation for Structured Pruning of a Large Language Model
72. ICLR 2026Reasoning Scaffolding: Distilling the Flow of Thought from LLMs
73. ICLR 2026Rectifying LLM Thought from Lens of Optimization
74. ICLR 2026RedSage: A Cybersecurity Generalist LLM
75. ICLR 2026RefineBench: Evaluating Refinement Capability in Language Models
76. ICLR 2026Rethinking Residual Errors in Compensation-based LLM Quantization
77. ICLR 2026Revisiting Parameter Server in LLM Post-Training
78. ICLR 2026Revisiting the Scaling Properties of Downstream Metrics in Large Language Model Training
79. ICLR 2026RL makes MLLMs see better than SFT
80. ICLR 2026RouterArena: An Open Platform for Comprehensive Comparison of LLM Routers
81. ICLR 2026Scheduling Your LLM Reinforcement Learning with Reasoning Trees
82. ICLR 2026SeedPrints: Fingerprints Can Even Tell Which Seed Your Large Language Model Was Trained From
83. ICLR 2026SelfReflect: Can LLMs Communicate Their Internal Answer Distribution?
84. ICLR 2026Semantic Regexes: Auto-Interpreting LLM Features with a Structured Language
85. ICLR 2026Semantic-Aware Diffusion LLM Inference With Adaptive Block Size
86. ICLR 2026SigLIP-HD by Fine-to-Coarse Supervision
87. ICLR 2026SK2Decompile: LLM-based Two-Phase Binary Decompilation from Skeleton to Skin
88. ICLR 2026Solving the Granularity Mismatch: Hierarchical Preference Learning for Long-Horizon LLM Agents
89. ICLR 2026SpareTrain: Fault-Tolerant LLM Training via Low-Cost Dual Modular Redundancy
90. ICLR 2026Stabilizing Policy Gradients for Sample-Efficient Reinforcement Learning in LLM Reasoning
91. ICLR 2026Taming the Fragility of KV Cache Eviction in LLM Inference
92. ICLR 2026Temperature as a Meta-Policy: Adaptive Temperature in LLM Reinforcement Learning
93. ICLR 2026Textual Bayes: Quantifying Uncertainty in LLM-Based Systems
94. ICLR 2026The Geometry of LLM Quantization: GPTQ as Babai's Nearest Plane Algorithm
95. ICLR 2026The Ideation-Execution Gap: Execution Outcomes of LLM-Generated versus Human Research Ideas
96. ICLR 2026The Mind's Transformer: Computational Neuroanatomy of LLM-Brain Alignment
97. ICLR 2026The Unseen Frontier: Pushing the Limits of LLM Sparsity with Surrogate-Free ADMM
98. ICLR 2026To Compress or Not? Pushing the Frontier of Lossless GenAI Model Weights Compression with Exponent Concentration
99. ICLR 2026Towards Multimodal Data-Driven Scientific Discovery Powered by LLM Agents
100. ICLR 2026TRAJECT-Bench:A Trajectory-Aware Benchmark for Evaluating Agentic Tool Use

常见问题

2023–2026 年四大会议一共有多少篇大模型基准论文？

共 544 篇，按年份分别是 2023 年 2 篇、2024 年 61 篇、2025 年 174 篇、2026 年 307 篇。

NeurIPS 2026 有多少篇大模型基准论文？

NeurIPS 2026 有 138 篇大模型基准论文，上一届是 92 篇。

ICLR 2026 有多少篇大模型基准论文？

ICLR 2026 有 103 篇大模型基准论文，上一届是 50 篇。

ICML 2026 有多少篇大模型基准论文？

ICML 2026 有 60 篇大模型基准论文，上一届是 24 篇。

CVPR 2026 有多少篇大模型基准论文？

CVPR 2026 有 6 篇大模型基准论文，上一届是 8 篇。

哪个会议的大模型基准论文最多？

NeurIPS，2023–2026 年共 256 篇。