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 688 篇，第 7 / 14 页

  1. NeurIPS 2026Idempotency Exposes Consistency Problems in Sparse Autoencoders
  2. NeurIPS 2026Integral Probability Boundaries
  3. NeurIPS 2026Internal Safety Collapse in Frontier Large Language Models
  4. NeurIPS 2026Jaguar: Fast Private CNN Inference with Power-of-Two Homomorphic Arithmetic
  5. NeurIPS 2026JailBound: A FOL-Guided Jailbreak Evaluation Framework for Revealing Safety Boundaries of LLMs
  6. NeurIPS 2026Learning Reach-Set Geometry for Tighter Probabilistic Neural Network Verification
  7. NeurIPS 2026LiSA: Lifelong Safety Adaptation via Conservative Policy Induction
  8. NeurIPS 2026LLM Alignment--Utility Asymmetry under Semantic-Preserving Transformations
  9. NeurIPS 2026LLM Rheology: Auditing Refusal Geometry in Aligned Language Models
  10. NeurIPS 2026Localizing Input Uncertainty Quantification for Large Language Models via Shapley Values
  11. NeurIPS 2026Loyalty Capture: Reporting Relationships and Structural Sycophancy in Frontier AI Models
  12. NeurIPS 2026Making Open-Source Text LLM Watermarks Durable Against Merging
  13. NeurIPS 2026MarkTune: Improving the Quality-Detectability Trade-off in Model-Embedded LLM Watermarking
  14. NeurIPS 2026MedMisBench: Measuring Epistemic Resilience of LLMs Under Misleading Medical Context
  15. NeurIPS 2026Metaphor Is Not All Attention Needs
  16. NeurIPS 2026Mitigating Memorization Where It Happens
  17. NeurIPS 2026Models Designed to Forget: Machine Unlearning via Key Deletion
  18. NeurIPS 2026Models That Know How Evaluations Are Designed Score Safer
  19. NeurIPS 2026Monitoring the Internal Monologue: Probe Trajectories Reveal Reasoning Dynamics
  20. NeurIPS 2026MOSAIC-Bench: Measuring Compositional Vulnerability Induction in Coding Agents
  21. NeurIPS 2026Multilingual Safety Alignment via Self-Distillation
  22. NeurIPS 2026Multistage Defer Trees for Hybrid Interpretability: If at First You Can't Succeed, Tree Again
  23. NeurIPS 2026MUTE: Multi-Level Alignment Uncoupling Against Talking-Head Exploitation for Voice Protection
  24. NeurIPS 2026On the Burden of Achieving Fairness in Conformal Prediction
  25. NeurIPS 2026Online Differentially Private Consistent Clustering
  26. NeurIPS 2026Optimal Rates for Differentially Private Hypothesis Testing with E-values
  27. NeurIPS 2026Optimizing Social Utility in Sequential Experiments
  28. NeurIPS 2026Parasite Features: Causal Abstraction Through Spurious Pathways
  29. NeurIPS 2026PDHFormer: Progressive Dual-Head Transformer for Behavioral Choice Prediction
  30. NeurIPS 2026Persona-Model Collapse in Emergent Misalignment
  31. NeurIPS 2026PrivateSeal: Low-Sensitivity Latent Directions for Diffusion-Resilient User-Specific Watermarking
  32. NeurIPS 2026PROACT-Agent: Progressive Runtime Oversight and Active Circuit-breaking for Real-Time Safety
  33. NeurIPS 2026Prompts to Proxies: Emulating Human Preferences via a Compact LLM Ensemble
  34. NeurIPS 2026Provable Explanations for Any-Order Neural Additive Models
  35. NeurIPS 2026ReasoningShield: Safety Moderation over Reasoning Traces of Large Reasoning Models
  36. NeurIPS 2026Rep2Text: Decoding Full Text from a Single LLM Token Representation
  37. NeurIPS 2026Retain-Neutral Surrogates for Min-Max Unlearning
  38. NeurIPS 2026Risks Create a Jagged Frontier of LLM Productivity Gains Across Computer Occupations
  39. NeurIPS 2026Rubric-Align: Safety Alignment through Dynamically Co-Evolving Rubrics
  40. NeurIPS 2026Safe Linear Bandits with Unknown Safety Gaps
  41. NeurIPS 2026Sample Transform Cost-Based Training-Free Hallucination Detector for Large Language Models
  42. NeurIPS 2026SAVeR$^2$: Reasoning-based Safety Alignment for Large Reasoning Models via Verifiable Rewards
  43. NeurIPS 2026Scalable Neural Safety Certification via Monotonicity
  44. NeurIPS 2026SciHazard: A Benchmark for Measuring Scientific Safety Risks with Decomposed Harm Scoring
  45. NeurIPS 2026SecureClaw: Clawing Back Control of LLM Agents
  46. NeurIPS 2026Selective Safety Steering via Value-Filtered Decoding
  47. NeurIPS 2026Steer-to-Detect: Probing Hidden Representations for Detection of LLM-Generated Texts
  48. NeurIPS 2026StereoTales: A Multilingual Framework for Open-Ended Stereotype Discovery in LLMs
  49. NeurIPS 2026T2V-AttnDisrupt: Inducing Hallucinations in LVLMs via Misrouting Visual Evidence Retrieval
  50. NeurIPS 2026TailGuard: Subgroup Tail Coverage Theory for Safe LLM Alignment

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