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 1,198 篇，第 18 / 24 页

  1. ICML 2025When, Where and Why to Average Weights?
  2. ICML 2024A connection between Tempering and Entropic Mirror Descent
  3. ICML 2024A Global Geometric Analysis of Maximal Coding Rate Reduction
  4. ICML 2024A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity
  5. ICML 2024A Minimaximalist Approach to Reinforcement Learning from Human Feedback
  6. ICML 2024A Neural-Preconditioned Poisson Solver for Mixed Dirichlet and Neumann Boundary Conditions
  7. ICML 2024A Unified Linear Programming Framework for Offline Reward Learning from Human Demonstrations and Feedback
  8. ICML 2024Accelerated Algorithms for Constrained Nonconvex-Nonconcave Min-Max Optimization and Comonotone Inclusion
  9. ICML 2024Active Preference Learning for Large Language Models
  10. ICML 2024Active Ranking and Matchmaking, with Perfect Matchings
  11. ICML 2024Advancing DRL Agents in Commercial Fighting Games: Training, Integration, and Agent-Human Alignment
  12. ICML 2024Adversarial Robustness Limits via Scaling-Law and Human-Alignment Studies
  13. ICML 2024Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast
  14. ICML 2024Agnostic Learning of Mixed Linear Regressions with EM and AM Algorithms
  15. ICML 2024AI Alignment with Changing and Influenceable Reward Functions
  16. ICML 2024AI Control: Improving Safety Despite Intentional Subversion
  17. ICML 2024Algorithm and Hardness for Dynamic Attention Maintenance in Large Language Models
  18. ICML 2024Aligning Transformers with Weisfeiler-Leman
  19. ICML 2024AlphaZero-Like Tree-Search can Guide Large Language Model Decoding and Training
  20. ICML 2024An Online Optimization Perspective on First-Order and Zero-Order Decentralized Nonsmooth Nonconvex Stochastic Optimization
  21. ICML 2024Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications
  22. ICML 2024AutoOS: Make Your OS More Powerful by Exploiting Large Language Models
  23. ICML 2024Balanced Resonate-and-Fire Neurons
  24. ICML 2024Balancing Similarity and Complementarity for Federated Learning
  25. ICML 2024Bootstrapping Fisher Market Equilibrium and First-Price Pacing Equilibrium
  26. ICML 2024BRAIn: Bayesian Reward-conditioned Amortized Inference for natural language generation from feedback
  27. ICML 2024Bridging Mini-Batch and Asymptotic Analysis in Contrastive Learning: From InfoNCE to Kernel-Based Losses
  28. ICML 2024Byzantine-Robust Federated Learning: Impact of Client Subsampling and Local Updates
  29. ICML 2024Can AI Assistants Know What They Don't Know?
  30. ICML 2024Can Machines Learn the True Probabilities?
  31. ICML 2024CARTE: Pretraining and Transfer for Tabular Learning
  32. ICML 2024Catapults in SGD: spikes in the training loss and their impact on generalization through feature learning
  33. ICML 2024Causal Bandits: The Pareto Optimal Frontier of Adaptivity, a Reduction to Linear Bandits, and Limitations around Unknown Marginals
  34. ICML 2024Characterizing Large Language Model Geometry Helps Solve Toxicity Detection and Generation
  35. ICML 2024Classification under Nuisance Parameters and Generalized Label Shift in Likelihood-Free Inference
  36. ICML 2024CogBench: a large language model walks into a psychology lab
  37. ICML 2024COLD-Attack: Jailbreaking LLMs with Stealthiness and Controllability
  38. ICML 2024Compressing Large Language Models by Joint Sparsification and Quantization
  39. ICML 2024Conformalized Survival Distributions: A Generic Post-Process to Increase Calibration
  40. ICML 2024Contrastive Predict-and-Search for Mixed Integer Linear Programs
  41. ICML 2024Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation
  42. ICML 2024Convergence and Complexity Guarantee for Inexact First-order Riemannian Optimization Algorithms
  43. ICML 2024Convergence and Trade-Offs in Riemannian Gradient Descent and Riemannian Proximal Point
  44. ICML 2024Convex and Bilevel Optimization for Neural-Symbolic Inference and Learning
  45. ICML 2024Criterion Collapse and Loss Distribution Control
  46. ICML 2024DecisionNCE: Embodied Multimodal Representations via Implicit Preference Learning
  47. ICML 2024Decoding-time Realignment of Language Models
  48. ICML 2024Decomposable Submodular Maximization in Federated Setting
  49. ICML 2024Deconstructing the Goldilocks Zone of Neural Network Initialization
  50. ICML 2024Decoupling Learning and Decision-Making: Breaking the $\mathcal{O}(\sqrt{T})$ Barrier in Online Resource Allocation with First-Order 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