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 277 篇，第 3 / 6 页

  1. NeurIPS 2026TailGuard: Subgroup Tail Coverage Theory for Safe LLM Alignment
  2. NeurIPS 2026Tatemae: Detecting Alignment Faking via Tool Selection in LLMs
  3. NeurIPS 2026Tcell: Mitigating Harmful Fine-tuning for Large Language Models via Gradient Alignment
  4. NeurIPS 2026Temporal Gradient Inversion for Private Trajectory Reconstruction in Embodied Reinforcement Learning
  5. NeurIPS 2026TensorCommitments: A Lightweight Verifiable Inference for Large Language Models
  6. NeurIPS 2026The Algorithm Is Not the Behavior: Learned Priors Override Look-Ahead in a Chess-Playing Neural Network
  7. NeurIPS 2026The Alignment Tax Concentrates in Output Projections
  8. NeurIPS 2026The Geometric Inductive Bias of Grokking: Bypassing Phase Transitions via Architectural Topology
  9. NeurIPS 2026The Geometric Wall: Manifold Structure Predicts Layerwise Sparse Autoencoder Scaling Laws
  10. NeurIPS 2026The Geometry of Alignment Collapse: When Fine-Tuning Breaks Safety
  11. NeurIPS 2026Tight PAC-Bayes Generalisation Guarantees for Large Language Model Safety Monitoring
  12. NeurIPS 2026Towards Fair Graph Generation Without Sensitive Attribute
  13. NeurIPS 2026Tracing the Cascade: A Topology-Aware Evaluation Framework for Scientific Agent Hallucinations
  14. NeurIPS 2026Training ML Models with Predictable Failures
  15. NeurIPS 2026TrustMod-SM: A Multi-Axis Benchmark for Evaluating Trustworthiness of LLMs in Social Media Content Moderation
  16. NeurIPS 2026Unsupervised Unlearnable Segmentation via Semantic Structure Disruption
  17. NeurIPS 2026Untrusted Content Masking for Web Agents with Security Guarantees
  18. NeurIPS 2026UpSafe℃: Upcycling for Controllable Safety in Large Language Models
  19. NeurIPS 2026Values Are Not Single Labels: Distributional Value Profiling Across Groups and Contexts
  20. NeurIPS 2026When Poison Meets Structure: Topology-based Defense against Poisoning Attack on Graph-based Retrieval-Augmented Generation
  21. NeurIPS 2026Where Should Society Draw the Line? A Social Choice Approach to Collective Consent
  22. NeurIPS 2025AdaptDel: Adaptable Deletion Rate Randomized Smoothing for Certified Robustness
  23. NeurIPS 2025Adaptive Defense against Harmful Fine-Tuning for Large Language Models via Bayesian Data Scheduler
  24. NeurIPS 2025Additive Models Explained: A Computational Complexity Approach
  25. NeurIPS 2025Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency
  26. NeurIPS 2025Alignment of Large Language Models with Constrained Learning
  27. NeurIPS 2025Analogy-based Multi-Turn Jailbreak against Large Language Models
  28. NeurIPS 2025Backdoor Cleaning without External Guidance in MLLM Fine-tuning
  29. NeurIPS 2025Boosting the Uniqueness of Neural Networks Fingerprints with Informative Triggers
  30. NeurIPS 2025BridgePure: Limited Protection Leakage Can Break Black-Box Data Protection
  31. NeurIPS 2025C-SafeGen: Certified Safe LLM Generation with Claim-Based Streaming Guardrails
  32. NeurIPS 2025CHiQPM: Calibrated Hierarchical Interpretable Image Classification
  33. NeurIPS 2025Collective Bargaining in the Information Economy Can Address AI-Driven Power Concentration
  34. NeurIPS 2025Combining Cost Constrained Runtime Monitors for AI Safety
  35. NeurIPS 2025Conformal Arbitrage: Risk-Controlled Balancing of Competing Objectives in Language Models
  36. NeurIPS 2025Contimask: Explaining Irregular Time Series via Perturbations in Continuous Time
  37. NeurIPS 2025CoT Red-Handed: Stress Testing Chain-of-Thought Monitoring
  38. NeurIPS 2025CTRL-ALT-DECEIT Sabotage Evaluations for Automated AI R&D
  39. NeurIPS 2025Detecting High-Stakes Interactions with Activation Probes
  40. NeurIPS 2025Detoxifying Large Language Models via Autoregressive Reward Guided Representation Editing
  41. NeurIPS 2025Direct Alignment with Heterogeneous Preferences
  42. NeurIPS 2025Discovering Opinion Intervals from Conflicts in Signed Graphs
  43. NeurIPS 2025Distillation Robustifies Unlearning
  44. NeurIPS 2025Does Representation Guarantee Welfare?
  45. NeurIPS 2025DualOptim: Enhancing Efficacy and Stability in Machine Unlearning with Dual Optimizers
  46. NeurIPS 2025Dynamic Risk Assessments for Offensive Cybersecurity Agents
  47. NeurIPS 2025Erasing Conceptual Knowledge from Language Models
  48. NeurIPS 2025ErrorTrace: A Black-Box Traceability Mechanism Based on Model Family Error Space
  49. NeurIPS 2025Fair Deepfake Detectors Can Generalize
  50. NeurIPS 2025Fairshare Data Pricing via Data Valuation for Large Language Models

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