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 688 篇，第 13 / 14 页

601. NeurIPS 2025SafeVid: Toward Safety Aligned Video Large Multimodal Models
602. NeurIPS 2025SAFEx: Analyzing Vulnerabilities of MoE-Based LLMs via Stable Safety-critical Expert Identification
603. NeurIPS 2025SAGE-Eval: Evaluating LLMs for Systematic Generalizations of Safety Facts
604. NeurIPS 2025Sample-efficient Learning of Concepts with Theoretical Guarantees: from Data to Concepts without Interventions
605. NeurIPS 2025Scalable Fingerprinting of Large Language Models
606. NeurIPS 2025Scalable Valuation of Human Feedback through Provably Robust Model Alignment
607. NeurIPS 2025Scaling Laws For Scalable Oversight
608. NeurIPS 2025Semantic Representation Attack against Aligned Large Language Models
609. NeurIPS 2025SHAP Meets Tensor Networks: Provably Tractable Explanations with Parallelism
610. NeurIPS 2025SHAP values via sparse Fourier representation
611. NeurIPS 2025SHAP zero Explains Biological Sequence Models with Near-zero Marginal Cost for Future Queries
612. NeurIPS 2025Shape it Up! Restoring LLM Safety during Finetuning
613. NeurIPS 2025Smoothed Differentiation Efficiently Mitigates Shattered Gradients in Explanations
614. NeurIPS 2025SpEx: A Spectral Approach to Explainable Clustering
615. NeurIPS 2025Spurious-Aware Prototype Refinement for Reliable Out-of-Distribution Detection
616. NeurIPS 2025STEAD: Robust Provably Secure Linguistic Steganography with Diffusion Language Model
617. NeurIPS 2025StegoZip: Enhancing Linguistic Steganography Payload in Practice with Large Language Models
618. NeurIPS 2025Strategic Classification with Non-Linear Classifiers
619. NeurIPS 2025Strategic Cost Selection in Participatory Budgeting
620. NeurIPS 2025Strategic Hypothesis Testing
621. NeurIPS 2025TAI3: Testing Agent Integrity in Interpreting User Intent
622. NeurIPS 2025Taming Hyperparameter Sensitivity in Data Attribution: Practical Selection Without Costly Retraining
623. NeurIPS 2025Taught Well Learned Ill: Towards Distillation-conditional Backdoor Attack
624. NeurIPS 2025The Future Unmarked: Watermark Removal in AI-Generated Images via Next-Frame Prediction
625. NeurIPS 2025The VLLM Safety Paradox: Dual Ease in Jailbreak Attack and Defense
626. NeurIPS 2025Towards Resilient Safety-driven Unlearning for Diffusion Models against Downstream Fine-tuning
627. NeurIPS 2025Towards Understanding Safety Alignment: A Mechanistic Perspective from Safety Neurons
628. NeurIPS 2025Transformer Key-Value Memories Are Nearly as Interpretable as Sparse Autoencoders
629. NeurIPS 2025Uncovering a Universal Abstract Algorithm for Modular Addition in Neural Networks
630. NeurIPS 2025Unlearning-Aware Minimization
631. NeurIPS 2025Unmasking Puppeteers: Leveraging Biometric Leakage to Expose Impersonation in AI-Based Videoconferencing
632. NeurIPS 2025Unveiling Environmental Sensitivity of Individual Gains in Influence Maximization
633. NeurIPS 2025Vector Database Watermarking
634. NeurIPS 2025Versatile Transferable Unlearnable Example Generator
635. NeurIPS 2025Video-SafetyBench: A Benchmark for Safety Evaluation of Video LVLMs
636. NeurIPS 2025Virus Infection Attack on LLMs: Your Poisoning Can Spread "VIA" Synthetic Data
637. NeurIPS 2025VLMs can Aggregate Scattered Training Patches
638. NeurIPS 2025What is Your Data Worth to GPT? LLM-Scale Data Valuation with Influence Functions
639. NeurIPS 2025What Really is a Member? Discrediting Membership Inference via Poisoning
640. NeurIPS 2025Whose View of Safety? A Deep DIVE Dataset for Pluralistic Alignment of Text-to-Image Models
641. NeurIPS 2025Why Do Some Language Models Fake Alignment While Others Don't?
642. NeurIPS 2025Wisdom is Knowing What not to Say: Hallucination-Free LLMs Unlearning via Attention Shifting
643. NeurIPS 2025World Models Should Prioritize the Unification of Physical and Social Dynamics
644. CVPR 2024Digital Life Project: Autonomous 3D Characters with Social Intelligence
645. ICLR 2024Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!
646. ICLR 2024Gaining Wisdom from Setbacks: Aligning Large Language Models via Mistake Analysis
647. ICLR 2024GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher
648. ICLR 2024Multilingual Jailbreak Challenges in Large Language Models
649. ICLR 2024Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions
650. ICLR 2024Social-Transmotion: Promptable Human Trajectory Prediction

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。