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 60 / 92 页

  1. NeurIPS 2025FRAM: Frobenius-Regularized Assignment Matching with Mixed-Precision Computing
  2. NeurIPS 2025FraPPE: Fast and Efficient Preference-Based Pure Exploration
  3. NeurIPS 2025From Judgment to Interference: Early Stopping LLM Harmful Outputs via Streaming Content Monitoring
  4. NeurIPS 2025From Noise to Narrative: Tracing the Origins of Hallucinations in Transformers
  5. NeurIPS 2025Fundamental Limitations in Pointwise Defences of LLM Finetuning APIs
  6. NeurIPS 2025GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs
  7. NeurIPS 2025Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel
  8. NeurIPS 2025Generalizing while preserving monotonicity in comparison-based preference learning models
  9. NeurIPS 2025Generative RLHF-V: Learning Principles from Multi-modal Human Preference
  10. NeurIPS 2025GMV: A Unified and Efficient Graph Multi-View Learning Framework
  11. NeurIPS 2025Gradient-Weight Alignment as a Train-Time Proxy for Generalization in Classification Tasks
  12. NeurIPS 2025Graph Alignment via Birkhoff Relaxation
  13. NeurIPS 2025GraSS: Scalable Data Attribution with Gradient Sparsification and Sparse Projection
  14. NeurIPS 2025Greedy Sampling Is Provably Efficient For RLHF
  15. NeurIPS 2025GRIFFIN: Effective Token Alignment for Faster Speculative Decoding
  16. NeurIPS 2025GSAlign: Geometric and Semantic Alignment Network for Aerial-Ground Person Re-Identification
  17. NeurIPS 2025GUARDIAN: Safeguarding LLM Multi-Agent Collaborations with Temporal Graph Modeling
  18. NeurIPS 2025Guiding Cross-Modal Representations with MLLM Priors via Preference Alignment
  19. NeurIPS 2025Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization
  20. NeurIPS 2025Harnessing Feature Resonance under Arbitrary Target Alignment for Out-of-Distribution Node Detection
  21. NeurIPS 2025HelpSteer3-Preference: Open Human-Annotated Preference Data across Diverse Tasks and Languages
  22. NeurIPS 2025Hierarchical Fine-grained Preference Optimization for Physically Plausible Video Generation
  23. NeurIPS 2025Hierarchical Frequency Tagging Probe (HFTP): A Unified Approach to Investigate Syntactic Structure Representations in Large Language Models and the Human Brain
  24. NeurIPS 2025How Does Label Noise Gradient Descent Improve Generalization in the Low SNR Regime?
  25. NeurIPS 2025Human-assisted Robotic Policy Refinement via Action Preference Optimization
  26. NeurIPS 2025HybridNorm: Towards Stable and Efficient Transformer Training via Hybrid Normalization
  27. NeurIPS 2025HYPERION: Fine-Grained Hypersphere Alignment for Robust Federated Graph Learning
  28. NeurIPS 2025Imbalances in Neurosymbolic Learning: Characterization and Mitigating Strategies
  29. NeurIPS 2025Impact of Layer Norm on Memorization and Generalization in Transformers
  30. NeurIPS 2025Impartial Selection with Predictions
  31. NeurIPS 2025Implicit Bias of Spectral Descent and Muon on Multiclass Separable Data
  32. NeurIPS 2025Implicit Reward as the Bridge: A Unified View of SFT and DPO Connections
  33. NeurIPS 2025Improved Bounds for Swap Multicalibration and Swap Omniprediction
  34. NeurIPS 2025Improved Confidence Regions and Optimal Algorithms for Online and Offline Linear MNL Bandits
  35. NeurIPS 2025Improved Representation Steering for Language Models
  36. NeurIPS 2025Improving LLM General Preference Alignment via Optimistic Online Mirror Descent
  37. NeurIPS 2025Improving Model Representation and Reducing KV Cache via Skip Connections with First Value Heads
  38. NeurIPS 2025Improving Video Generation with Human Feedback
  39. NeurIPS 2025Incomplete Multi-view Clustering via Hierarchical Semantic Alignment and Cooperative Completion
  40. NeurIPS 2025Incomplete Multi-view Deep Clustering with Data Imputation and Alignment
  41. NeurIPS 2025Inference-time Alignment in Continuous Space
  42. NeurIPS 2025Inference-Time Personalized Alignment with a Few User Preference Queries
  43. NeurIPS 2025Inference-Time Reward Hacking in Large Language Models
  44. NeurIPS 2025Inference-Time Scaling for Flow Models via Stochastic Generation and Rollover Budget Forcing
  45. NeurIPS 2025Inference-Time Text-to-Video Alignment with Diffusion Latent Beam Search
  46. NeurIPS 2025InfiFPO: Implicit Model Fusion via Preference Optimization in Large Language Models
  47. NeurIPS 2025Information Retrieval Induced Safety Degradation in AI Agents
  48. NeurIPS 2025Information-Theoretic Reward Decomposition for Generalizable RLHF
  49. NeurIPS 2025Infrequent Exploration in Linear Bandits
  50. NeurIPS 2025Integrating Drug Substructures and Longitudinal Electronic Health Records for Personalized Drug Recommendation

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