社会层面 · safety论文趋势 2023–2026

主要结论

• 社会层面 · safety论文从 2023 年的 4 篇增长到 2026 年的 370 篇，三年间扩大到约 93 倍。
• ICML 是发社会层面 · safety论文最多的会议，2023–2026 年共 364 篇，占四个会议总量的 53%。
• 交叉方向里涨得最快的是对齐与安全：同时属于两个主题的论文从 2023 年的 0 篇增至 2026 年的 137 篇。
• 在「对齐与负责任 AI」的 5 个主题里，它的 2026 年论文数排第 3；同组增长最快的社会层面 · privacy是 6.6 倍，社会层面 · safety是 93 倍。

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

• NeurIPS 2026BarrierSteer: LLM Safety via Learning Barrier Steering
• ICML 2026Anti-Backdoor Coreset Selection via Cumulative Entropy
• ICLR 2026Bi-Criteria Metric Distortion
• CVPR 2026Omni-MMSI Toward Identity-attributed Social Interaction Understanding
• NeurIPS 2025Safety Pretraining: Toward the Next Generation of Safe AI
• ICML 2025$S^2$FGL: Spatial Spectral Federated Graph Learning
• ICLR 2025The Hidden Cost of Waiting for Accurate Predictions
• CVPR 2025Do We Really Need Curated Malicious Data for Safety Alignment in Multi-modal Large Language Models?
• NeurIPS 2024Quantifying the Bitter Lesson: How Safety Benchmarks Measure Capabilities Instead of Safety
• ICML 2024Optimally Improving Cooperative Learning in a Social Setting
• ICLR 2024The Devil is in the Neurons: Interpreting and Mitigating Social Biases in Language Models
• CVPR 2024Digital Life Project: Autonomous 3D Characters with Social Intelligence
• NeurIPS 2023Strategyproof Voting under Correlated Beliefs
• ICML 2023Adaptive Coordination in Social Embodied Rearrangement
• NeurIPS 2026Beyond Feature Disruption: Boundary-Diverting Unlearnable Examples against Linear Probing
• ICML 2026AutoControl Arena: Synthesizing Executable Test Environments for Frontier AI Risk Evaluation
• ICLR 2026PRISON: Unmasking the Criminal Potential of Large Language Models
• CVPR 2026Spot The Ball A Benchmark for Visual Social Inference
• NeurIPS 2025SAFEx: Analyzing Vulnerabilities of MoE-Based LLMs via Stable Safety-critical Expert Identification
• ICML 2025Adversaries Can Misuse Combinations of Safe Models

全部社会层面 · safety论文

共 226 篇，第 1 / 5 页

1. ICML 2026A Geometric Analysis of Small-sized Language Model Hallucinations
2. ICML 2026A New Framework for Cybersecurity Refusals in AI Agents
3. ICML 2026A Positive Case for Faithfulness: Explanations Help Predict Model Behavior
4. ICML 2026A Unified Framework for Diffusion Model Unlearning with f-Divergence
5. ICML 2026A-MemGuard: A Proactive Defense Framework For LLM-Based Agent Memory
6. ICML 2026ABC-Bench: An Agentic Bio-Capabilities Benchmark for Biosecurity
7. ICML 2026Acoustic Interference: A New Paradigm Weaponizing Acoustic Latent Semantic for Universal Jailbreak against Large Audio Language Models
8. ICML 2026Agentic Confidence Calibration
9. ICML 2026AIR: Improving Agent Safety through Incident Response
10. ICML 2026Algorithmic Recourse of In-Context Learning for Tabular Data
11. ICML 2026Ambiguous Strategic Classification
12. ICML 2026An Odd Estimator for Shapley Values
13. ICML 2026Anchored Decoding: Provably Reducing Copyright Risk for Any Language Model
14. ICML 2026Anti-Backdoor Coreset Selection via Cumulative Entropy
15. ICML 2026Architecture Matters for Multi-Agent Security
16. ICML 2026Are Your Agents Upward Deceivers?
17. ICML 2026At the Edge of Understanding: Sparse Autoencoders Trace The Limits of Transformer Generalization
18. ICML 2026AutoControl Arena: Synthesizing Executable Test Environments for Frontier AI Risk Evaluation
19. ICML 2026Automatically Finding Reward Model Biases
20. ICML 2026Belief Dynamics Reveal the Dual Nature of In-Context Learning and Activation Steering
21. ICML 2026Beyond Additive Decompositions: Interpretability Through Separability
22. ICML 2026Beyond External Monitors: Enhancing Transparency of Large Language Models for Easier Monitoring
23. ICML 2026Beyond Pixels: Mining Compressed Domain Artifacts for Efficient AI-Generated Video Detection
24. ICML 2026Biases in the Blind Spot: Detecting What LLMs Fail to Mention
25. ICML 2026BlueCodeAgent: A Blue Teaming Agent Powered by Automated Red Teaming for CodeGen AI
26. ICML 2026Broadening the Backdoor Basin: Understanding LLM Backdoors Collapse and Making Backdoors Persistent
27. ICML 2026Building Reliable Long-Form Generation via Hallucination Rejection Sampling
28. ICML 2026Calibrating Conservatism for Scalable Oversight
29. ICML 2026CapBencher: Give Your LLM Benchmark a Built-in Alarm for Test-Set Overfitting
30. ICML 2026Certified Circuits: Stability Guarantees for Mechanistic Circuits
31. ICML 2026Collaborative Disagreement Resolution for Scalable Oversight
32. ICML 2026Collaborative Threshold Watermarking
33. ICML 2026Concept Concentration for Faithful Representation Intervention
34. ICML 2026Consistency Training Can Entrench Misalignment
35. ICML 2026Constitutional Black-Box Monitoring for Scheming in LLM Agents
36. ICML 2026Contrastive Reasoning Alignment: Reinforcement Learning from Hidden Representations
37. ICML 2026CoopEval: Benchmarking Cooperation-Sustaining Mechanisms and LLM Agents in Social Dilemmas
38. ICML 2026Copyright-Bench: Agentic Evaluation of Copyright Law Compliance
39. ICML 2026CORE: Conflict-Oriented Reasoning for General Multimodal Manipulation Detection
40. ICML 2026Corrigibility Transformation: Constructing Goals That Accept Updates
41. ICML 2026CoT is Not the Chain of Truth: An Empirical Internal Analysis of Reasoning LLMs for Fake News Generation
42. ICML 2026D-Judge: Disrupting Multi-Turn Jailbreaks using Semantics-Preserving Output Rewriting
43. ICML 2026DDIM Inversion as a Perturbation Amplifier: Breaking Mimicry Protection via Reconstruction Error Minimization
44. ICML 2026Debate with Images: Detecting Deceptive Behaviors in Multimodal Large Language Models
45. ICML 2026Decomposing Query-Key Feature Interactions Using Contrastive Covariances
46. ICML 2026Detecting Fluent Optimization-Based Adversarial Prompts via Sequential Entropy Changes
47. ICML 2026Diagnosing the Reliability of LLM-as-a-Judge via Item Response Theory
48. ICML 2026Dimensional Collapse in Transformer Attention Outputs: A Challenge for Sparse Dictionary Learning
49. ICML 2026Discovering Implicit Large Language Model Alignment Objectives
50. ICML 2026Dismantling Pathological Shortcuts: A Causal Framework for Faithful LVLM Decoding

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。