社会层面 · safety论文趋势 2023–2026

AI 社会议题：safety

统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 688 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

社会层面 · safety论文数，按年份和会议
年份NeurIPSICLRICMLCVPR合计
202320204
202423116141
2025131111301273
2026121202263370

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

全部社会层面 · safety论文

年份全部2023202420252026
会议全部NeurIPSICLRICMLCVPR

共 273 篇，第 5 / 6 页

  1. NeurIPS 2025Monoculture or Multiplicity: Which Is It?
  2. NeurIPS 2025Noise Injection Reveals Hidden Capabilities of Sandbagging Language Models
  3. NeurIPS 2025One Head to Rule Them All: Amplifying LVLM Safety through a Single Critical Attention Head
  4. NeurIPS 2025OOD Detection with Relative Angles
  5. NeurIPS 2025OrdShap: Feature Position Importance for Sequential Black-Box Models
  6. NeurIPS 2025OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents
  7. NeurIPS 2025Overcoming Sparsity Artifacts in Crosscoders to Interpret Chat-Tuning
  8. NeurIPS 2025Panacea: Mitigating Harmful Fine-tuning for Large Language Models via Post-fine-tuning Perturbation
  9. NeurIPS 2025Performative Validity of Recourse Explanations
  10. NeurIPS 2025PolyGuard: Massive Multi-Domain Safety Policy-Grounded Guardrail Dataset
  11. NeurIPS 2025Predicting the Performance of Black-box Language Models with Follow-up Queries
  12. NeurIPS 2025Preference Learning with Lie Detectors can Induce Honesty or Evasion
  13. NeurIPS 2025Probabilistic Stability Guarantees for Feature Attributions
  14. NeurIPS 2025Provable Gradient Editing of Deep Neural Networks
  15. NeurIPS 2025Provably Efficient RL under Episode-Wise Safety in Constrained MDPs with Linear Function Approximation
  16. NeurIPS 2025Rectifying Shortcut Behaviors in Preference-based Reward Learning
  17. NeurIPS 2025Redefining Experts: Interpretable Decomposition of Language Models for Toxicity Mitigation
  18. NeurIPS 2025Refusal Direction is Universal Across Safety-Aligned Languages
  19. NeurIPS 2025Regional Explanations: Bridging Local and Global Variable Importance
  20. NeurIPS 2025Regression Trees Know Calculus
  21. NeurIPS 2025Reliably detecting model failures in deployment without labels
  22. NeurIPS 2025Representational Difference Explanations
  23. NeurIPS 2025Residual Stream Analysis of Overfitting And Structural Disruptions
  24. NeurIPS 2025RiOSWorld: Benchmarking the Risk of Multimodal Computer-Use Agents
  25. NeurIPS 2025Risk Management for Mitigating Benchmark Failure Modes: BenchRisk
  26. NeurIPS 2025Robust Explanations of Graph Neural Networks via Graph Curvatures
  27. NeurIPS 2025Safe RLHF-V: Safe Reinforcement Learning from Multi-modal Human Feedback
  28. NeurIPS 2025SAFEPATH: Preventing Harmful Reasoning in Chain-of-Thought via Early Alignment
  29. NeurIPS 2025Safety Depth in Large Language Models: A Markov Chain Perspective
  30. NeurIPS 2025Safety Pretraining: Toward the Next Generation of Safe AI
  31. NeurIPS 2025SafeVid: Toward Safety Aligned Video Large Multimodal Models
  32. NeurIPS 2025SAFEx: Analyzing Vulnerabilities of MoE-Based LLMs via Stable Safety-critical Expert Identification
  33. NeurIPS 2025SAGE-Eval: Evaluating LLMs for Systematic Generalizations of Safety Facts
  34. NeurIPS 2025Sample-efficient Learning of Concepts with Theoretical Guarantees: from Data to Concepts without Interventions
  35. NeurIPS 2025Scalable Fingerprinting of Large Language Models
  36. NeurIPS 2025Scalable Valuation of Human Feedback through Provably Robust Model Alignment
  37. NeurIPS 2025Scaling Laws For Scalable Oversight
  38. NeurIPS 2025Semantic Representation Attack against Aligned Large Language Models
  39. NeurIPS 2025SHAP Meets Tensor Networks: Provably Tractable Explanations with Parallelism
  40. NeurIPS 2025SHAP values via sparse Fourier representation
  41. NeurIPS 2025SHAP zero Explains Biological Sequence Models with Near-zero Marginal Cost for Future Queries
  42. NeurIPS 2025Shape it Up! Restoring LLM Safety during Finetuning
  43. NeurIPS 2025Smoothed Differentiation Efficiently Mitigates Shattered Gradients in Explanations
  44. NeurIPS 2025SpEx: A Spectral Approach to Explainable Clustering
  45. NeurIPS 2025Spurious-Aware Prototype Refinement for Reliable Out-of-Distribution Detection
  46. NeurIPS 2025STEAD: Robust Provably Secure Linguistic Steganography with Diffusion Language Model
  47. NeurIPS 2025StegoZip: Enhancing Linguistic Steganography Payload in Practice with Large Language Models
  48. NeurIPS 2025Strategic Classification with Non-Linear Classifiers
  49. NeurIPS 2025Strategic Cost Selection in Participatory Budgeting
  50. NeurIPS 2025Strategic Hypothesis Testing

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 查社会层面 · safety在 ICML 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": "social-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社会层面 · safety论文？
共 688 篇，按年份分别是 2023 年 4 篇、2024 年 41 篇、2025 年 273 篇、2026 年 370 篇。
NeurIPS 2026 有多少篇社会层面 · safety论文？
NeurIPS 2026 有 121 篇社会层面 · safety论文，上一届是 131 篇。
ICLR 2026 有多少篇社会层面 · safety论文？
ICLR 2026 有 20 篇社会层面 · safety论文，上一届是 11 篇。
ICML 2026 有多少篇社会层面 · safety论文？
ICML 2026 有 226 篇社会层面 · safety论文，上一届是 130 篇。
CVPR 2026 有多少篇社会层面 · safety论文？
CVPR 2026 有 3 篇社会层面 · safety论文，上一届是 1 篇。
哪个会议的社会层面 · safety论文最多？
ICML，2023–2026 年共 364 篇。

「对齐与负责任 AI」下的其他主题