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 11 / 14 页

  1. ICML 2025Towards Black-Box Membership Inference Attack for Diffusion Models
  2. ICML 2025Towards flexible perception with visual memory
  3. ICML 2025Towards Global-level Mechanistic Interpretability: A Perspective of Modular Circuits of Large Language Models
  4. ICML 2025TRUST-VLM: Thorough Red-Teaming for Uncovering Safety Threats in Vision-Language Models
  5. ICML 2025Trustworthy Machine Learning through Data-Specific Indistinguishability
  6. ICML 2025Uncertainty-Based Extensible Codebook for Discrete Federated Learning in Heterogeneous Data Silos
  7. ICML 2025Validating Mechanistic Interpretations: An Axiomatic Approach
  8. ICML 2025Weak-to-Strong Jailbreaking on Large Language Models
  9. ICML 2025What makes an Ensemble (Un) Interpretable?
  10. ICML 2025Which Attention Heads Matter for In-Context Learning?
  11. ICML 2025WMarkGPT: Watermarked Image Understanding via Multimodal Large Language Models
  12. ICML 2025X-Hacking: The Threat of Misguided AutoML
  13. NeurIPS 2025AdaptDel: Adaptable Deletion Rate Randomized Smoothing for Certified Robustness
  14. NeurIPS 2025Adaptive Defense against Harmful Fine-Tuning for Large Language Models via Bayesian Data Scheduler
  15. NeurIPS 2025Additive Models Explained: A Computational Complexity Approach
  16. NeurIPS 2025Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency
  17. NeurIPS 2025Alignment of Large Language Models with Constrained Learning
  18. NeurIPS 2025Analogy-based Multi-Turn Jailbreak against Large Language Models
  19. NeurIPS 2025Backdoor Cleaning without External Guidance in MLLM Fine-tuning
  20. NeurIPS 2025Boosting the Uniqueness of Neural Networks Fingerprints with Informative Triggers
  21. NeurIPS 2025BridgePure: Limited Protection Leakage Can Break Black-Box Data Protection
  22. NeurIPS 2025C-SafeGen: Certified Safe LLM Generation with Claim-Based Streaming Guardrails
  23. NeurIPS 2025CHiQPM: Calibrated Hierarchical Interpretable Image Classification
  24. NeurIPS 2025Collective Bargaining in the Information Economy Can Address AI-Driven Power Concentration
  25. NeurIPS 2025Combining Cost Constrained Runtime Monitors for AI Safety
  26. NeurIPS 2025Conformal Arbitrage: Risk-Controlled Balancing of Competing Objectives in Language Models
  27. NeurIPS 2025Contimask: Explaining Irregular Time Series via Perturbations in Continuous Time
  28. NeurIPS 2025CoT Red-Handed: Stress Testing Chain-of-Thought Monitoring
  29. NeurIPS 2025CTRL-ALT-DECEIT Sabotage Evaluations for Automated AI R&D
  30. NeurIPS 2025Detecting High-Stakes Interactions with Activation Probes
  31. NeurIPS 2025Detoxifying Large Language Models via Autoregressive Reward Guided Representation Editing
  32. NeurIPS 2025Direct Alignment with Heterogeneous Preferences
  33. NeurIPS 2025Discovering Opinion Intervals from Conflicts in Signed Graphs
  34. NeurIPS 2025Distillation Robustifies Unlearning
  35. NeurIPS 2025Does Representation Guarantee Welfare?
  36. NeurIPS 2025DualOptim: Enhancing Efficacy and Stability in Machine Unlearning with Dual Optimizers
  37. NeurIPS 2025Dynamic Risk Assessments for Offensive Cybersecurity Agents
  38. NeurIPS 2025Erasing Conceptual Knowledge from Language Models
  39. NeurIPS 2025ErrorTrace: A Black-Box Traceability Mechanism Based on Model Family Error Space
  40. NeurIPS 2025Fair Deepfake Detectors Can Generalize
  41. NeurIPS 2025Fairshare Data Pricing via Data Valuation for Large Language Models
  42. NeurIPS 2025Faithful Group Shapley Value
  43. NeurIPS 2025FALCON: Fine-grained Activation Manipulation by Contrastive Orthogonal Unalignment for Large Language Model
  44. NeurIPS 2025Final-Model-Only Data Attribution with a Unifying View of Gradient-Based Methods
  45. NeurIPS 2025From Black-box to Causal-box: Towards Building More Interpretable Models
  46. NeurIPS 2025From Judgment to Interference: Early Stopping LLM Harmful Outputs via Streaming Content Monitoring
  47. NeurIPS 2025From Noise to Narrative: Tracing the Origins of Hallucinations in Transformers
  48. NeurIPS 2025From Specificity to Generality: Revisiting Generalizable Artifacts in Detecting Face Deepfakes
  49. NeurIPS 2025Fundamental Limitations in Pointwise Defences of LLM Finetuning APIs
  50. NeurIPS 2025GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs

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