对齐与安全论文趋势 2023–2026

主要结论

• 对齐与安全论文从 2023 年的 438 篇增长到 2026 年的 1,882 篇，三年间扩大到约 4.3 倍。
• NeurIPS 是发对齐与安全论文最多的会议，2023–2026 年共 1,421 篇，占四个会议总量的 31%。
• 交叉方向里涨得最快的是社会层面 · safety：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 1；同组增长最快的社会层面 · safety是 93 倍，对齐与安全是 4.3 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 2026$\gamma$-weakly $\theta$-up-concavity: A Unified Framework for Non-Convex Optimization Beyond DR-Submodular and OSS Functions
• ICML 2026AICrypto: Evaluating Cryptography Capabilities of Large Language Models
• ICLR 2026A Fair Bayesian Inference through Matched Gibbs Posterior
• CVPR 2026Advancing Cancer Prognosis with Hierarchical Fusion of Genomic Proteomic and
• NeurIPS 2025$\Psi$-Sampler: Initial Particle Sampling for SMC-Based Inference-Time Reward Alignment in Score Models
• ICML 2025A Checks-and-Balances Framework for Context-Aware Ethical AI Alignment
• ICLR 2025$\sigma$-zero: Gradient-based Optimization of $\ell_0$-norm Adversarial Examples
• CVPR 2025AesthetiQ: Enhancing Graphic Layout Design via Aesthetic-Aware Preference Alignment of Multi-modal Large Language Models
• NeurIPS 2024A survey and benchmark of high-dimensional Bayesian optimization of discrete sequences
• ICML 2024A Global Geometric Analysis of Maximal Coding Rate Reduction
• ICLR 2024A unique M-pattern for micro-expression spotting in long videos
• CVPR 2024Decompose-and-Compose: A Compositional Approach to Mitigating Spurious Correlation
• NeurIPS 2023(S)GD over Diagonal Linear Networks: Implicit bias, Large Stepsizes and Edge of Stability
• ICML 2023A Robust Optimisation Perspective on Counterexample-Guided Repair of Neural Networks
• ICLR 2023Aligning Model and Macaque Inferior Temporal Cortex Representations Improves Model-to-Human Behavioral Alignment and Adversarial Robustness
• CVPR 2023Adjustment and Alignment for Unbiased Open Set Domain Adaptation
• NeurIPS 2026$\mathcal{P}$Torch: Narrowing the Gap Between Projection and Gradient-Based Learning
• ICML 2026Alignment Pretraining: AI Discourse Causes Self-Fulfilling (Mis)alignment
• ICLR 2026A2D: Any-Order, Any-Step Safety Alignment for Diffusion Language Models
• CVPR 2026AlignPose Generalizable 6D Pose Estimation via Multi-view Feature-metric Alignment

全部对齐与安全论文

共 1,363 篇，第 22 / 28 页

1051. NeurIPS 2025ErrorTrace: A Black-Box Traceability Mechanism Based on Model Family Error Space
1052. NeurIPS 2025ESCA: Enabling Seamless Codec Avatar Execution through Algorithm and Hardware Co-Optimization for Virtual Reality
1053. NeurIPS 2025EVOREFUSE: Evolutionary Prompt Optimization for Evaluation and Mitigation of LLM Over-Refusal to Pseudo-Malicious Instructions
1054. NeurIPS 2025Explainable Reinforcement Learning from Human Feedback to Improve Alignment
1055. NeurIPS 2025Explaining and Mitigating Crosslingual Tokenizer Inequities
1056. NeurIPS 2025Explaining the Law of Supply and Demand via Online Learning
1057. NeurIPS 2025Exploring Data Scaling Trends and Effects in Reinforcement Learning from Human Feedback
1058. NeurIPS 2025Fact-R1: Towards Explainable Video Misinformation Detection with Deep Reasoning
1059. NeurIPS 2025Failure Prediction at Runtime for Generative Robot Policies
1060. NeurIPS 2025Fair Minimum Labeling: Efficient Temporal Network Activations for Reachability and Equity
1061. NeurIPS 2025Fast Computation and Optimization for Opinion-Based Quantities of Friedkin-Johnsen Model
1062. NeurIPS 2025FastJAM: a Fast Joint Alignment Model for Images
1063. NeurIPS 2025FedFree: Breaking Knowledge-sharing Barriers through Layer-wise Alignment in Heterogeneous Federated Learning
1064. NeurIPS 2025FedWMSAM: Fast and Flat Federated Learning via Weighted Momentum and Sharpness-Aware Minimization
1065. NeurIPS 2025Finding and Reactivating Post-Trained LLMs' Hidden Safety Mechanisms
1066. NeurIPS 2025Fine-Grained Preference Optimization Improves Spatial Reasoning in VLMs
1067. NeurIPS 2025Fine-Tuning Discrete Diffusion Models with Policy Gradient Methods
1068. NeurIPS 2025Fire360: A Benchmark for Robust Perception and Episodic Memory in Degraded 360° Firefighting Video
1069. NeurIPS 2025FRAM: Frobenius-Regularized Assignment Matching with Mixed-Precision Computing
1070. NeurIPS 2025FraPPE: Fast and Efficient Preference-Based Pure Exploration
1071. NeurIPS 2025From Judgment to Interference: Early Stopping LLM Harmful Outputs via Streaming Content Monitoring
1072. NeurIPS 2025From Noise to Narrative: Tracing the Origins of Hallucinations in Transformers
1073. NeurIPS 2025Fundamental Limitations in Pointwise Defences of LLM Finetuning APIs
1074. NeurIPS 2025GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs
1075. NeurIPS 2025Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel
1076. NeurIPS 2025Generalizing while preserving monotonicity in comparison-based preference learning models
1077. NeurIPS 2025Generative RLHF-V: Learning Principles from Multi-modal Human Preference
1078. NeurIPS 2025GMV: A Unified and Efficient Graph Multi-View Learning Framework
1079. NeurIPS 2025Gradient-Weight Alignment as a Train-Time Proxy for Generalization in Classification Tasks
1080. NeurIPS 2025Graph Alignment via Birkhoff Relaxation
1081. NeurIPS 2025GraSS: Scalable Data Attribution with Gradient Sparsification and Sparse Projection
1082. NeurIPS 2025Greedy Sampling Is Provably Efficient For RLHF
1083. NeurIPS 2025GRIFFIN: Effective Token Alignment for Faster Speculative Decoding
1084. NeurIPS 2025GSAlign: Geometric and Semantic Alignment Network for Aerial-Ground Person Re-Identification
1085. NeurIPS 2025GUARDIAN: Safeguarding LLM Multi-Agent Collaborations with Temporal Graph Modeling
1086. NeurIPS 2025Guiding Cross-Modal Representations with MLLM Priors via Preference Alignment
1087. NeurIPS 2025Handling Label Noise via Instance-Level Difficulty Modeling and Dynamic Optimization
1088. NeurIPS 2025Harnessing Feature Resonance under Arbitrary Target Alignment for Out-of-Distribution Node Detection
1089. NeurIPS 2025HelpSteer3-Preference: Open Human-Annotated Preference Data across Diverse Tasks and Languages
1090. NeurIPS 2025Hierarchical Fine-grained Preference Optimization for Physically Plausible Video Generation
1091. NeurIPS 2025Hierarchical Frequency Tagging Probe (HFTP): A Unified Approach to Investigate Syntactic Structure Representations in Large Language Models and the Human Brain
1092. NeurIPS 2025How Does Label Noise Gradient Descent Improve Generalization in the Low SNR Regime?
1093. NeurIPS 2025Human-assisted Robotic Policy Refinement via Action Preference Optimization
1094. NeurIPS 2025HybridNorm: Towards Stable and Efficient Transformer Training via Hybrid Normalization
1095. NeurIPS 2025HYPERION: Fine-Grained Hypersphere Alignment for Robust Federated Graph Learning
1096. NeurIPS 2025Imbalances in Neurosymbolic Learning: Characterization and Mitigating Strategies
1097. NeurIPS 2025Impact of Layer Norm on Memorization and Generalization in Transformers
1098. NeurIPS 2025Impartial Selection with Predictions
1099. NeurIPS 2025Implicit Bias of Spectral Descent and Muon on Multiclass Separable Data
1100. NeurIPS 2025Implicit Reward as the Bridge: A Unified View of SFT and DPO Connections

常见问题

2023–2026 年四大会议一共有多少篇对齐与安全论文？

共 4,558 篇，按年份分别是 2023 年 438 篇、2024 年 875 篇、2025 年 1,363 篇、2026 年 1,882 篇。

NeurIPS 2026 有多少篇对齐与安全论文？

NeurIPS 2026 有 480 篇对齐与安全论文，上一届是 455 篇。

ICLR 2026 有多少篇对齐与安全论文？

ICLR 2026 有 623 篇对齐与安全论文，上一届是 473 篇。

ICML 2026 有多少篇对齐与安全论文？

ICML 2026 有 551 篇对齐与安全论文，上一届是 300 篇。

CVPR 2026 有多少篇对齐与安全论文？

CVPR 2026 有 228 篇对齐与安全论文，上一届是 135 篇。

哪个会议的对齐与安全论文最多？

NeurIPS，2023–2026 年共 1,421 篇。