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 623 篇，第 5 / 13 页

  1. ICLR 2026Exponential-Wrapped Mechanisms: Differential Privacy on Hadamard Manifolds Made Practical
  2. ICLR 2026Exposing and Defending the Achilles' Heel of Video Mixture-of-Experts
  3. ICLR 2026Expressiveness of Multi-Neuron Convex Relaxations in Neural Network Certification
  4. ICLR 2026Fair Classification by Direct Intervention on Operating Characteristics
  5. ICLR 2026Fair Conformal Classification via Learning Representation-Based Groups
  6. ICLR 2026Fair Decision Utility in Human-AI Collaboration: Interpretable Confidence Adjustment for Humans with Cognitive Disparities
  7. ICLR 2026Fair Graph Machine Learning under Adversarial Missingness Processes
  8. ICLR 2026Fair in Mind, Fair in Action? A Synchronous Benchmark for Understanding and Generation in UMLLMs
  9. ICLR 2026Fair Policy Aggregation from Standard Policy Optimization
  10. ICLR 2026Fairness via Independence: A General Regularization Framework for Machine Learning
  11. ICLR 2026Faithful Bi-Directional Model Steering via Distribution Matching and Distributed Interchange Interventions
  12. ICLR 2026FaLW: A Forgetting-aware Loss Reweighting for Long-tailed Unlearning
  13. ICLR 2026Fast and Interpretable Protein Substructure Alignment via Optimal Transport
  14. ICLR 2026Fast and Stable Riemannian Metrics on SPD Manifolds via Cholesky Product Geometry
  15. ICLR 2026Federated Learning of Quantile Inference under Local Differential Privacy
  16. ICLR 2026Federated Learning with Profile Mapping under Distribution Shifts and Drifts
  17. ICLR 2026FedMC: Federated Manifold Calibration
  18. ICLR 2026Fewer Weights, More Problems: A Practical Attack on LLM Pruning
  19. ICLR 2026Fine-Grained Iterative Adversarial Attacks with Limited Computation Budget
  20. ICLR 2026Fine-tuning Done Right in Model Editing
  21. ICLR 2026Fingerprinting Deep Neural Networks for Ownership Protection: An Analytical Approach
  22. ICLR 2026FLARE: Fully Integration of Vision-Language Representations for Deep Cross-Modal Understanding
  23. ICLR 2026Flattery, Fluff, and Fog: Diagnosing and Mitigating Idiosyncratic Biases in Preference Models
  24. ICLR 2026FlexHiNM-GP: Flexible Hierarchical Pruning via Region Allocation and Channel Permutation
  25. ICLR 2026FLoRG: Federated Fine-tuning with Low-rank Gram Matrices and Procrustes Alignment
  26. ICLR 2026FlowSearcher: Synthesizing Memory-Guided Agentic Workflows for Web Information Seeking
  27. ICLR 2026Fluent Alignment with Disfluent Judges: Post-training for lower-resource languages
  28. ICLR 2026Follow-Your-Preference: Towards Preference-Aligned Image Inpainting
  29. ICLR 2026Formalising Human-in-the-Loop: Computational Reductions, Failure Modes, and Legal-Moral Responsibility
  30. ICLR 2026From Evaluation to Defense: Advancing Safety in Video Large Language Models
  31. ICLR 2026From Natural Alignment to Conditional Controllability in Multimodal Dialogue
  32. ICLR 2026From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning
  33. ICLR 2026From Utterance to Vividity: Training Expressive Subtitle Translation LLM via Adaptive Local Preference Optimization
  34. ICLR 2026FrontierCO: Real-World and Large-Scale Evaluation of Machine Learning Solvers for Combinatorial Optimization
  35. ICLR 2026FSPO: Few-Shot Optimization of Synthetic Preferences Effectively Personalizes to Real Users
  36. ICLR 2026Full-Graph vs. Mini-Batch Training: Comprehensive Analysis from a Batch Size and Fan-Out Size Perspective
  37. ICLR 2026GAPrune: Gradient-Alignment Pruning for Domain-Aware Embeddings
  38. ICLR 2026Gaussian certified unlearning in high dimensions: A hypothesis testing approach
  39. ICLR 2026GAVEL: Towards Rule-Based Safety through Activation Monitoring
  40. ICLR 2026Gelato: Graph Edit Distance via Autoregressive Neural Combinatorial Optimization
  41. ICLR 2026General Exploratory Bonus for Optimistic Exploration in RLHF
  42. ICLR 2026Generative Bayesian Optimization: Generative Models as Acquisition Functions
  43. ICLR 2026Generative Value Conflicts Reveal LLM Priorities
  44. ICLR 2026Ghost in the Cloud: Your Geo-Distributed Large Language Models Training is Easily Manipulated
  45. ICLR 2026GhostEI-Bench: Do Mobile Agent Resilience to Environmental Injection in Dynamic On-Device Environments?
  46. ICLR 2026GLASS Flows: Efficient Inference for Reward Alignment of Flow and Diffusion Models
  47. ICLR 2026GNN Explanations that do not Explain and How to find Them
  48. ICLR 2026Good allocations from bad estimates
  49. ICLR 2026GOT-Edit: Geometry-Aware Generic Object Tracking via Online Model Editing
  50. ICLR 2026GPTailor: Large Language Model Pruning Through Layer Cutting and Stitching

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