模型即裁判论文趋势 2023–2026

自动评测

统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 401 篇被归入该主题的论文。

主要结论

每年论文数

各会议论文数

按年份和会议统计

模型即裁判论文数，按年份和会议
年份NeurIPSICLRICMLCVPR合计
202351208
202427129048
20255337333126
20268862645219

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

全部模型即裁判论文

年份全部2023202420252026
会议全部NeurIPSICLRICMLCVPR

共 401 篇，第 5 / 9 页

  1. NeurIPS 2026Strong Teacher Not Needed? On Distillation in LLM Pretraining
  2. NeurIPS 2026TextSeal: A Localized LLM Watermark for Provenance & Distillation Protection
  3. NeurIPS 2026The Geometry of Forgetting: Temporal Knowledge Drift as an Independent Axis in LLM Representations
  4. NeurIPS 2026The Missing Corpus: Infinite Ground Truth for File-Grounded LLM Evaluation
  5. NeurIPS 2026The Ruler and the Judge: Benchmark-Conditional Evaluation of LLM-as-a-Judge
  6. NeurIPS 2026Too Smart to Teach: The Articulability Ceiling in Language Models
  7. NeurIPS 2026Towards Reliable LLM Evaluation: Correcting the Winner’s Curse in Adaptive Benchmarking
  8. NeurIPS 2026Treat Domain-Specific Languages as Design Variables in LLM Agents
  9. NeurIPS 2026Trust, but Don’t Verify: Epistemic Blind Spots in LLM Source Evaluation
  10. NeurIPS 2026TwinRouterBench: Fast Static and Live Dynamic Evaluation for Realistic Agentic LLM Routing
  11. NeurIPS 2026Valid Best-Model Identification for LLM Evaluation via Low-Rank Factorization
  12. NeurIPS 2026Verifiable LLM-Guided Focal SMT Solving for Quantified Arrays
  13. NeurIPS 2026Verified-Source Authority Is Not Generic Sycophancy: Cue-Family Decomposition of LLM Compliance
  14. NeurIPS 2026VEX-Bench: Benchmarking Verification Complexity of LLM-Generated Misinformation
  15. NeurIPS 2026What Claims Do LLM Benchmark Scores Support?
  16. NeurIPS 2026What Should Remain After Forgetting? Rethinking LLM Unlearning as Predictive Posterior Correction
  17. NeurIPS 2026When Empathy Misses the Goal: A Benchmark for Goal Displacement in LLM Advice
  18. NeurIPS 2026When Expert Disagreement Hurts: Auditing Prestige-Sensitive Revision in LLM Decision Pipelines
  19. NeurIPS 2026When LLM Routers Overpay: Strong-Model Over-Selection under Loose Budgets
  20. CVPR 2025ASHiTA: Automatic Scene-grounded HIerarchical Task Analysis
  21. CVPR 2025LLaVA-Critic: Learning to Evaluate Multimodal Models
  22. CVPR 2025Period-LLM: Extending the Periodic Capability of Multimodal Large Language Model
  23. ICLR 2025A Statistical Framework for Ranking LLM-based Chatbots
  24. ICLR 2025AgentSquare: Automatic LLM Agent Search in Modular Design Space
  25. ICLR 2025An Empirical Analysis of Uncertainty in Large Language Model Evaluations
  26. ICLR 2025BadJudge: Backdoor Vulnerabilities of LLM-As-A-Judge
  27. ICLR 2025Better than Your Teacher: LLM Agents that learn from Privileged AI Feedback
  28. ICLR 2025Beyond correlation: The impact of human uncertainty in measuring the effectiveness of automatic evaluation and LLM-as-a-judge
  29. ICLR 2025Catastrophic Failure of LLM Unlearning via Quantization
  30. ICLR 2025Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Rates
  31. ICLR 2025Combatting Dimensional Collapse in LLM Pre-Training Data via Submodular File Selection
  32. ICLR 2025Controlled LLM Decoding via Discrete Auto-regressive Biasing
  33. ICLR 2025Diversity Empowers Intelligence: Integrating Expertise of Software Engineering Agents
  34. ICLR 2025Do LLMs have Consistent Values?
  35. ICLR 2025GraphEval: A Lightweight Graph-Based LLM Framework for Idea Evaluation
  36. ICLR 2025HShare: Fast LLM Decoding by Hierarchical Key-Value Sharing
  37. ICLR 2025Judge Decoding: Faster Speculative Sampling Requires Going Beyond Model Alignment
  38. ICLR 2025JudgeBench: A Benchmark for Evaluating LLM-Based Judges
  39. ICLR 2025Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge
  40. ICLR 2025Language Model Alignment in Multilingual Trolley Problems
  41. ICLR 2025Language Models Are Implicitly Continuous
  42. ICLR 2025Latent Space Chain-of-Embedding Enables Output-free LLM Self-Evaluation
  43. ICLR 2025Learning LLM-as-a-Judge for Preference Alignment
  44. ICLR 2025Limits to scalable evaluation at the frontier: LLM as judge won’t beat twice the data
  45. ICLR 2025LiveBench: A Challenging, Contamination-Limited LLM Benchmark
  46. ICLR 2025On Calibration of LLM-based Guard Models for Reliable Content Moderation
  47. ICLR 2025Pangea: A Fully Open Multilingual Multimodal LLM for 39 Languages
  48. ICLR 2025Perplexed by Perplexity: Perplexity-Based Data Pruning With Small Reference Models
  49. ICLR 2025Progressive Mixed-Precision Decoding for Efficient LLM Inference
  50. ICLR 2025RevisEval: Improving LLM-as-a-Judge via Response-Adapted References

在 Claude、Cursor 里直接问这个趋势

接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。

可以直接复制这句话问 AI

用 paperlist 查模型即裁判在 NeurIPS 2023–2026 的论文趋势，哪一年涨得最快？

AI 会调用的工具：get_topic_stats {"topic": "llm-as-judge"}

两分钟接入 Paperlist MCP（Claude Code、Cursor、Codex、ChatGPT）

常见问题

2023–2026 年四大会议一共有多少篇模型即裁判论文？
共 401 篇，按年份分别是 2023 年 8 篇、2024 年 48 篇、2025 年 126 篇、2026 年 219 篇。
NeurIPS 2026 有多少篇模型即裁判论文？
NeurIPS 2026 有 88 篇模型即裁判论文，上一届是 53 篇。
ICLR 2026 有多少篇模型即裁判论文？
ICLR 2026 有 62 篇模型即裁判论文，上一届是 37 篇。
ICML 2026 有多少篇模型即裁判论文？
ICML 2026 有 64 篇模型即裁判论文，上一届是 33 篇。
CVPR 2026 有多少篇模型即裁判论文？
CVPR 2026 有 5 篇模型即裁判论文，上一届是 3 篇。
哪个会议的模型即裁判论文最多？
NeurIPS，2023–2026 年共 173 篇。

「评测与基准」下的其他主题