对齐与安全论文趋势 2023–2026

RLHF、偏好优化、红队与安全对齐

统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 4,558 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

对齐与安全论文数，按年份和会议
年份NeurIPSICLRICMLCVPR合计
202313887101112438
2024348156246125875
20254554733001351,363
20264806235512281,882

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

全部对齐与安全论文

年份全部2023202420252026
会议全部NeurIPSICLRICMLCVPR

共 4,558 篇，第 49 / 92 页

  1. ICLR 2025Self-Boosting Large Language Models with Synthetic Preference Data
  2. ICLR 2025SELF-EVOLVED REWARD LEARNING FOR LLMS
  3. ICLR 2025Self-Improvement in Language Models: The Sharpening Mechanism
  4. ICLR 2025Self-Improving Robust Preference Optimization
  5. ICLR 2025Self-Play Preference Optimization for Language Model Alignment
  6. ICLR 2025Self-play with Execution Feedback: Improving Instruction-following Capabilities of Large Language Models
  7. ICLR 2025Sensitivity Verification for Additive Decision Tree Ensembles
  8. ICLR 2025SeRA: Self-Reviewing and Alignment of LLMs using Implicit Reward Margins
  9. ICLR 2025Severing Spurious Correlations with Data Pruning
  10. ICLR 2025Shape as Line Segments: Accurate and Flexible Implicit Surface Representation
  11. ICLR 2025Shh, don't say that! Domain Certification in LLMs
  12. ICLR 2025SimPER: A Minimalist Approach to Preference Alignment without Hyperparameters
  13. ICLR 2025Simple yet Effective Incomplete Multi-view Clustering: Similarity-level Imputation and Intra-view Hybrid-group Prototype Construction
  14. ICLR 2025SimulPL: Aligning Human Preferences in Simultaneous Machine Translation
  15. ICLR 2025SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal
  16. ICLR 2025Sparse components distinguish visual pathways & their alignment to neural networks
  17. ICLR 2025Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment
  18. ICLR 2025Spurious Forgetting in Continual Learning of Language Models
  19. ICLR 2025Stealthy Shield Defense: A Conditional Mutual Information-Based Approach against Black-Box Model Inversion Attacks
  20. ICLR 2025Steering Masked Discrete Diffusion Models via Discrete Denoising Posterior Prediction
  21. ICLR 2025Strategic Classification With Externalities
  22. ICLR 2025Strong Preferences Affect the Robustness of Preference Models and Value Alignment
  23. ICLR 2025Style Outweighs Substance: Failure Modes of LLM Judges in Alignment Benchmarking
  24. ICLR 2025Super(ficial)-alignment: Strong Models May Deceive Weak Models in Weak-to-Strong Generalization
  25. ICLR 2025SuperCorrect: Advancing Small LLM Reasoning with Thought Template Distillation and Self-Correction
  26. ICLR 2025Support is All You Need for Certified VAE Training
  27. ICLR 2025Surgical, Cheap, and Flexible: Mitigating False Refusal in Language Models via Single Vector Ablation
  28. ICLR 2025Tailoring Mixup to Data for Calibration
  29. ICLR 2025Taming Overconfidence in LLMs: Reward Calibration in RLHF
  30. ICLR 2025Tamper-Resistant Safeguards for Open-Weight LLMs
  31. ICLR 2025TASAR: Transfer-based Attack on Skeletal Action Recognition
  32. ICLR 2025Test-time Adaptation for Regression by Subspace Alignment
  33. ICLR 2025Test-time Alignment of Diffusion Models without Reward Over-optimization
  34. ICLR 2025The Crucial Role of Samplers in Online Direct Preference Optimization
  35. ICLR 2025The Effectiveness of Curvature-Based Rewiring and the Role of Hyperparameters in GNNs Revisited
  36. ICLR 2025The Geometry of Categorical and Hierarchical Concepts in Large Language Models
  37. ICLR 2025The Hidden Cost of Waiting for Accurate Predictions
  38. ICLR 2025The Last Iterate Advantage: Empirical Auditing and Principled Heuristic Analysis of Differentially Private SGD
  39. ICLR 2025The Optimization Landscape of SGD Across the Feature Learning Strength
  40. ICLR 2025The Utility and Complexity of In- and Out-of-Distribution Machine Unlearning
  41. ICLR 2025Tighter Privacy Auditing of DP-SGD in the Hidden State Threat Model
  42. ICLR 2025TIS-DPO: Token-level Importance Sampling for Direct Preference Optimization With Estimated Weights
  43. ICLR 2025To Tackle Adversarial Transferability: A Novel Ensemble Training Method with Fourier Transformation
  44. ICLR 2025To Trust or Not to Trust? Enhancing Large Language Models' Situated Faithfulness to External Contexts
  45. ICLR 2025TODO: Enhancing LLM Alignment with Ternary Preferences
  46. ICLR 2025Toward Guidance-Free AR Visual Generation via Condition Contrastive Alignment
  47. ICLR 2025Towards a learning theory of representation alignment
  48. ICLR 2025Towards Certification of Uncertainty Calibration under Adversarial Attacks
  49. ICLR 2025Towards counterfactual fairness through auxiliary variables
  50. ICLR 2025Towards Effective Evaluations and Comparisons for LLM Unlearning Methods

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 查对齐与安全在 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": "alignment-and-safety"}

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对齐与安全论文？
共 4,558 篇，按年份分别是 2023 年 438 篇、2024 年 875 篇、2025 年 1,363 篇、2026 年 1,882 篇。
NeurIPS 2026 有多少篇对齐与安全论文？
NeurIPS 2026 有 480 篇对齐与安全论文，上一届是 455 篇。
ICLR 2026 有多少篇对齐与安全论文？
ICLR 2026 有 623 篇对齐与安全论文，上一届是 473 篇。
ICML 2026 有多少篇对齐与安全论文？
ICML 2026 有 551 篇对齐与安全论文，上一届是 300 篇。
CVPR 2026 有多少篇对齐与安全论文？
CVPR 2026 有 228 篇对齐与安全论文，上一届是 135 篇。
哪个会议的对齐与安全论文最多？
NeurIPS，2023–2026 年共 1,421 篇。

「对齐与负责任 AI」下的其他主题