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 52 / 92 页

  1. ICML 2025DAMA: Data- and Model-aware Alignment of Multi-modal LLMs
  2. ICML 2025DANCE: Dual Unbiased Expansion with Group-acquired Alignment for Out-of-distribution Graph Fairness Learning
  3. ICML 2025Deep Linear Network Training Dynamics from Random Initialization: Data, Width, Depth, and Hyperparameter Transfer
  4. ICML 2025Defending LVLMs Against Vision Attacks Through Partial-Perception Supervision
  5. ICML 2025Design Considerations in Offline Preference-based RL
  6. ICML 2025Differentiable Quadratic Optimization For the Maximum Independent Set Problem
  7. ICML 2025Direct Density Ratio Optimization: A Statistically Consistent Approach to Aligning Large Language Models
  8. ICML 2025Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator
  9. ICML 2025Discrete and Continuous Difference of Submodular Minimization
  10. ICML 2025Discriminative Finetuning of Generative Large Language Models without Reward Models and Human Preference Data
  11. ICML 2025Distributed Retraction-Free and Communication-Efficient Optimization on the Stiefel Manifold
  12. ICML 2025Diverging Preferences: When do Annotators Disagree and do Models Know?
  13. ICML 2025Don't Restart, Just Reuse: Reoptimizing MILPs with Dynamic Parameters
  14. ICML 2025DPO Meets PPO: Reinforced Token Optimization for RLHF
  15. ICML 2025DreamDPO: Aligning Text-to-3D Generation with Human Preferences via Direct Preference Optimization
  16. ICML 2025Edge-Colored Clustering in Hypergraphs: Beyond Minimizing Unsatisfied Edges
  17. ICML 2025Eigen Analysis of Conjugate Kernel and Neural Tangent Kernel
  18. ICML 2025Eliciting Language Model Behaviors with Investigator Agents
  19. ICML 2025ELITE: Enhanced Language-Image Toxicity Evaluation for Safety
  20. ICML 2025ELMO : Efficiency via Low-precision and Peak Memory Optimization in Large Output Spaces
  21. ICML 2025Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs
  22. ICML 2025Energy-Based Preference Model Offers Better Offline Alignment than the Bradley-Terry Preference Model
  23. ICML 2025Enhancing Ligand Validity and Affinity in Structure-Based Drug Design with Multi-Reward Optimization
  24. ICML 2025Enhancing Parallelism in Decentralized Stochastic Convex Optimization
  25. ICML 2025Enhancing Spectral GNNs: From Topology and Perturbation Perspectives
  26. ICML 2025EquivaMap: Leveraging LLMs for Automatic Equivalence Checking of Optimization Formulations
  27. ICML 2025EraseAnything: Enabling Concept Erasure in Rectified Flow Transformers
  28. ICML 2025Evolving Prompts In-Context: An Open-ended, Self-replicating Perspective
  29. ICML 2025Exact risk curves of signSGD in High-Dimensions: quantifying preconditioning and noise-compression effects
  30. ICML 2025Explicit Preference Optimization: No Need for an Implicit Reward Model
  31. ICML 2025Faster Global Minimum Cut with Predictions
  32. ICML 2025Features are fate: a theory of transfer learning in high-dimensional regression
  33. ICML 2025FlipAttack: Jailbreak LLMs via Flipping
  34. ICML 2025Flow Matching for Few-Trial Neural Adaptation with Stable Latent Dynamics
  35. ICML 2025Functional Alignment Can Mislead: Examining Model Stitching
  36. ICML 2025Fusing Reward and Dueling Feedback in Stochastic Bandits
  37. ICML 2025Gandalf the Red: Adaptive Security for LLMs
  38. ICML 2025Generalizing from SIMPLE to HARD Visual Reasoning: Can We Mitigate Modality Imbalance in VLMs?
  39. ICML 2025Global Convergence and Rich Feature Learning in $L$-Layer Infinite-Width Neural Networks under $\mu$ Parametrization
  40. ICML 2025Global Optimization with a Power-Transformed Objective and Gaussian Smoothing
  41. ICML 2025Gradient Descent Converges Arbitrarily Fast for Logistic Regression via Large and Adaptive Stepsizes
  42. ICML 2025GRAIL: Graph Edit Distance and Node Alignment using LLM-Generated Code
  43. ICML 2025GRAM: A Generative Foundation Reward Model for Reward Generalization
  44. ICML 2025Hardware and Software Platform Inference
  45. ICML 2025Hierarchical Graph Tokenization for Molecule-Language Alignment
  46. ICML 2025How to set AdamW's weight decay as you scale model and dataset size
  47. ICML 2025HPS: Hard Preference Sampling for Human Preference Alignment
  48. ICML 2025Human-Aligned Image Models Improve Visual Decoding from the Brain
  49. ICML 2025ICLShield: Exploring and Mitigating In-Context Learning Backdoor Attacks
  50. ICML 2025Implicit Language Models are RNNs: Balancing Parallelization and Expressivity

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