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 273 篇，第 1 / 6 页

1. CVPR 2025Do We Really Need Curated Malicious Data for Safety Alignment in Multi-modal Large Language Models?
2. ICLR 2025A Black Swan Hypothesis: The Role of Human Irrationality in AI Safety
3. ICLR 2025Breach By A Thousand Leaks: Unsafe Information Leakage in 'Safe' AI Responses
4. ICLR 2025Controllable Safety Alignment: Inference-Time Adaptation to Diverse Safety Requirements
5. ICLR 2025Does Safety Training of LLMs Generalize to Semantically Related Natural Prompts?
6. ICLR 2025Durable Quantization Conditioned Misalignment Attack on Large Language Models
7. ICLR 2025Jailbreak Antidote: Runtime Safety-Utility Balance via Sparse Representation Adjustment in Large Language Models
8. ICLR 2025On the Role of Attention Heads in Large Language Model Safety
9. ICLR 2025Programming Refusal with Conditional Activation Steering
10. ICLR 2025Safety Alignment Should be Made More Than Just a Few Tokens Deep
11. ICLR 2025Safety Layers in Aligned Large Language Models: The Key to LLM Security
12. ICLR 2025The Hidden Cost of Waiting for Accurate Predictions
13. ICML 2025(How) Can Transformers Predict Pseudo-Random Numbers?
14. ICML 2025$S^2$FGL: Spatial Spectral Federated Graph Learning
15. ICML 2025A Checks-and-Balances Framework for Context-Aware Ethical AI Alignment
16. ICML 2025A Lens into Interpretable Transformer Mistakes via Semantic Dependency
17. ICML 2025A Non-Asymptotic Convergent Analysis for Scored-Based Graph Generative Model via a System of Stochastic Differential Equations
18. ICML 2025A Versatile Influence Function for Data Attribution with Non-Decomposable Loss
19. ICML 2025Activation Space Interventions Can Be Transferred Between Large Language Models
20. ICML 2025Adversarial Reasoning at Jailbreaking Time
21. ICML 2025Adversaries Can Misuse Combinations of Safe Models
22. ICML 2025Algorithmic Recourse for Long-Term Improvement
23. ICML 2025An Efficient Private GPT Never Autoregressively Decodes
24. ICML 2025An End-to-End Model for Logits-Based Large Language Models Watermarking
25. ICML 2025Antidote: Post-fine-tuning Safety Alignment for Large Language Models against Harmful Fine-tuning Attack
26. ICML 2025Assessing Safety Risks and Quantization-aware Safety Patching for Quantized Large Language Models
27. ICML 2025Automated Red Teaming with GOAT: the Generative Offensive Agent Tester
28. ICML 2025Automatically Interpreting Millions of Features in Large Language Models
29. ICML 2025Avoiding Leakage Poisoning: Concept Interventions Under Distribution Shifts
30. ICML 2025AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders
31. ICML 2025B-score: Detecting biases in large language models using response history
32. ICML 2025CASE-Bench: Context-Aware SafEty Benchmark for Large Language Models
33. ICML 2025Certified Unlearning for Neural Networks
34. ICML 2025Circumventing Backdoor Space via Weight Symmetry
35. ICML 2025Confidential Guardian: Cryptographically Prohibiting the Abuse of Model Abstention
36. ICML 2025Correlated Errors in Large Language Models
37. ICML 2025Detecting Strategic Deception with Linear Probes
38. ICML 2025DIS-CO: Discovering Copyrighted Content in VLMs Training Data
39. ICML 2025Discovering Spoofing Attempts on Language Model Watermarks
40. ICML 2025Disparate Conditional Prediction in Multiclass Classifiers
41. ICML 2025Do Not Mimic My Voice : Speaker Identity Unlearning for Zero-Shot Text-to-Speech
42. ICML 2025Efficient and Privacy-Preserving Soft Prompt Transfer for LLMs
43. ICML 2025ELITE: Enhanced Language-Image Toxicity Evaluation for Safety
44. ICML 2025Embedding Safety into RL: A New Take on Trust Region Methods
45. ICML 2025Emoji Attack: Enhancing Jailbreak Attacks Against Judge LLM Detection
46. ICML 2025EraseAnything: Enabling Concept Erasure in Rectified Flow Transformers
47. ICML 2025Evaluating Neuron Explanations: A Unified Framework with Sanity Checks
48. ICML 2025Explaining, Fast and Slow: Abstraction and Refinement of Provable Explanations
49. ICML 2025ExpProof : Operationalizing Explanations for Confidential Models with ZKPs
50. ICML 2025Extractive Structures Learned in Pretraining Enable Generalization on Finetuned Facts

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。