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 166 篇，第 3 / 4 页

101. ICLR 2026When LLMs get significantly worse: A statistical approach to detect model degradations
102. ICLR 2026WideSearch: Benchmarking Agentic Broad Info-Seeking
103. ICLR 2026WSM: Decay-Free Learning Rate Schedule via Checkpoint Merging for LLM Pre-training
104. ICLR 2025AI Sandbagging: Language Models can Strategically Underperform on Evaluations
105. ICLR 2025An Empirical Analysis of Uncertainty in Large Language Model Evaluations
106. ICLR 2025Better than Your Teacher: LLM Agents that learn from Privileged AI Feedback
107. ICLR 2025BingoGuard: LLM Content Moderation Tools with Risk Levels
108. ICLR 2025Catastrophic Failure of LLM Unlearning via Quantization
109. ICLR 2025Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates
110. ICLR 2025Combatting Dimensional Collapse in LLM Pre-Training Data via Submodular File Selection
111. ICLR 2025Controlled LLM Decoding via Discrete Auto-regressive Biasing
112. ICLR 2025Cut the Crap: An Economical Communication Pipeline for LLM-based Multi-Agent Systems
113. ICLR 2025Determine-Then-Ensemble: Necessity of Top-k Union for Large Language Model Ensembling
114. ICLR 2025Diversity Empowers Intelligence: Integrating Expertise of Software Engineering Agents
115. ICLR 2025Do LLMs have Consistent Values?
116. ICLR 2025Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMs
117. ICLR 2025DON’T STOP ME NOW: EMBEDDING BASED SCHEDULING FOR LLMS
118. ICLR 2025ELICIT: LLM Augmentation Via External In-context Capability
119. ICLR 2025Faster Cascades via Speculative Decoding
120. ICLR 2025ForecastBench: A Dynamic Benchmark of AI Forecasting Capabilities
121. ICLR 2025Fourier Head: Helping Large Language Models Learn Complex Probability Distributions
122. ICLR 2025From Commands to Prompts: LLM-based Semantic File System for AIOS
123. ICLR 2025GameArena: Evaluating LLM Reasoning through Live Computer Games
124. ICLR 2025Herald: A Natural Language Annotated Lean 4 Dataset
125. ICLR 2025HShare: Fast LLM Decoding by Hierarchical Key-Value Sharing
126. ICLR 2025Improving Large Language Model Planning with Action Sequence Similarity
127. ICLR 2025Instruct-SkillMix: A Powerful Pipeline for LLM Instruction Tuning
128. ICLR 2025Instructional Segment Embedding: Improving LLM Safety with Instruction Hierarchy
129. ICLR 2025Language Models Are Implicitly Continuous
130. ICLR 2025Large Language Models can Become Strong Self-Detoxifiers
131. ICLR 2025Law of the Weakest Link: Cross Capabilities of Large Language Models
132. ICLR 2025Lawma: The Power of Specialization for Legal Annotation
133. ICLR 2025LiveBench: A Challenging, Contamination-Limited LLM Benchmark
134. ICLR 2025LLM Unlearning via Loss Adjustment with Only Forget Data
135. ICLR 2025Logical Consistency of Large Language Models in Fact-Checking
136. ICLR 2025MM1.5: Methods, Analysis & Insights from Multimodal LLM Fine-tuning
137. ICLR 2025Omni-MATH: A Universal Olympiad Level Mathematic Benchmark for Large Language Models
138. ICLR 2025On Calibration of LLM-based Guard Models for Reliable Content Moderation
139. ICLR 2025Perplexed by Perplexity: Perplexity-Based Data Pruning With Small Reference Models
140. ICLR 2025PhyloLM: Inferring the Phylogeny of Large Language Models and Predicting their Performances in Benchmarks
141. ICLR 2025RevisEval: Improving LLM-as-a-Judge via Response-Adapted References
142. ICLR 2025Robotouille: An Asynchronous Planning Benchmark for LLM Agents
143. ICLR 2025SeedLM: Compressing LLM Weights into Seeds of Pseudo-Random Generators
144. ICLR 2025Selective Attention Improves Transformer
145. ICLR 2025SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal
146. ICLR 2025SpinQuant: LLM Quantization with Learned Rotations
147. ICLR 2025SqueezeAttention: 2D Management of KV-Cache in LLM Inference via Layer-wise Optimal Budget
148. ICLR 2025StringLLM: Understanding the String Processing Capability of Large Language Models
149. ICLR 2025Style Outweighs Substance: Failure Modes of LLM Judges in Alignment Benchmarking
150. ICLR 2025TODO: Enhancing LLM Alignment with Ternary Preferences

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。