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 103 篇，第 1 / 3 页

1. ICLR 2026<SO$G_k$>: One LLM Token for Explicit Graph Structural Understanding
2. ICLR 2026Activation Steering for LLM Alignment via a Unified ODE-Based Framework
3. ICLR 2026Adaptive Thinking: Large Language Models Know When to Think in Latent Space
4. ICLR 2026ASCIIEval: Benchmarking Models' Visual Perception in Text Strings via ASCII Art
5. ICLR 2026AutoSP: Unlocking Long-Context LLM Training Via Compiler-Based Sequence Parallelism
6. ICLR 2026Beyond Length: Quantifying Long-Range Information for Long-Context LLM Pretraining Data
7. ICLR 2026Calibrating Verbalized Confidence with Self-Generated Distractors
8. ICLR 2026Cancer-Myth: Evaluating Large Language Models on Patient Questions with False Presuppositions
9. ICLR 2026Capability-Based Scaling Trends for LLM-Based Red-Teaming
10. ICLR 2026CMPhysBench: A Benchmark for Evaluating Large Language Models in Condensed Matter Physics
11. ICLR 2026Common Corpus: The Largest Collection of Ethical Data for LLM Pre-Training
12. ICLR 2026Count Counts: Motivating Exploration in LLM Reasoning with Count-based Intrinsic Rewards
13. ICLR 2026Credit-Budgeted ICPC-Style Coding: When LLM Agents Must Pay for Every Decision
14. ICLR 2026DASH: Deterministic Attention Scheduling for High-throughput Reproducible LLM Training
15. ICLR 2026Data Selection for LLM Alignment Using Fine-Grained Preferences
16. ICLR 2026DeepResearch Bench: A Comprehensive Benchmark for Deep Research Agents
17. ICLR 2026Detecting Data Contamination in LLMs via In-Context Learning
18. ICLR 2026Discern Truth from Falsehood: Reducing Over-Refusal via Contrastive Refinement
19. ICLR 2026Discovering Novel LLM Experts via Task-Capability Coevolution
20. ICLR 2026Downgrade to Upgrade: Optimizer Simplification Enhances Robustness in LLM Unlearning
21. ICLR 2026Dr.LLM: Dynamic Layer Routing in LLMs
22. ICLR 2026EditBench: Evaluating LLM Abilities to Perform Real-World Instructed Code Edits
23. ICLR 2026Empowering LLM Tool Invocation with Tool-call Reward Model
24. ICLR 2026Enhancing Instruction Following of LLMs via Activation Steering with Dynamic Rejection
25. ICLR 2026Epistemic Uncertainty Quantification To Improve Decisions From Black-Box Models
26. ICLR 2026Expanding the Capability Frontier of LLM Agents with ZPD-Guided Data Synthesis
27. ICLR 2026ExpGuard: LLM Content Moderation in Specialized Domains
28. ICLR 2026Fast-dLLM v2: Efficient Block-Diffusion LLM
29. ICLR 2026From Static Benchmarks to Dynamic Protocol: Agent-Centric Text Anomaly Detection for Evaluating LLM Reasoning
30. ICLR 2026FutureX: An Advanced Live Benchmark for LLM Agents in Future Prediction
31. ICLR 2026Go-Browse: Training Web Agents with Structured Exploration
32. ICLR 2026Hey, That's My Model! Introducing Chain & Hash, An LLM Fingerprinting Technique
33. ICLR 2026How Do Transformers Learn to Associate Tokens: Gradient Leading Terms Bring Mechanistic Interpretability
34. ICLR 2026How Learning Rate Decay Wastes Your Best Data in Curriculum-Based LLM Pretraining
35. ICLR 2026How Stable is the Next Token? A Geometric View of LLM Prediction Stability
36. ICLR 2026Improving LLM-based Global Optimization with Search Space Partitioning
37. ICLR 2026In Agents We Trust, but Who Do Agents Trust? Latent Preferences Steer LLM Generations
38. ICLR 2026Jet Expansions: Restructuring LLM Computation for Model Inspection
39. ICLR 2026KL-Regularized Reinforcement Learning is Designed to Mode Collapse
40. ICLR 2026KV-Cache Transform Coding for Compact Storage in LLM Inference
41. ICLR 2026LLM as an Algorithmist: Enhancing Anomaly Detectors via Programmatic Synthesis
42. ICLR 2026LLM Pretraining with Continuous Concepts
43. ICLR 2026LLM Unlearning with LLM Beliefs
44. ICLR 2026LLM-as-a-Prophet: Understanding Predictive Intelligence with Prophet Arena
45. ICLR 2026Log Probability Tracking of LLM APIs
46. ICLR 2026LS-Merge: Merging Language Models in Latent Space
47. ICLR 2026LycheeDecode: Accelerating Long-Context LLM Inference via Hybrid-Head Sparse Decoding
48. ICLR 2026Making Slow Thinking Faster: Compressing LLM Chain-of-Thought via Step Entropy
49. ICLR 2026MCP Security Bench (MSB): Benchmarking Attacks Against Model Context Protocol in LLM Agents
50. ICLR 2026MCP-Bench: Benchmarking Tool-Using LLM Agents with Complex Real-World Tasks via MCP Servers

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。