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 156 篇，第 3 / 4 页

101. ICLR 2024Quantifying and Enhancing Multi-modal Robustness with Modality Preference
102. ICLR 2024Query-Dependent Prompt Evaluation and Optimization with Offline Inverse RL
103. ICLR 2024Quick-Tune: Quickly Learning Which Pretrained Model to Finetune and How
104. ICLR 2024RAIN: Your Language Models Can Align Themselves without Finetuning
105. ICLR 2024RECOMBINER: Robust and Enhanced Compression with Bayesian Implicit Neural Representations
106. ICLR 2024Respect the model: Fine-grained and Robust Explanation with Sharing Ratio Decomposition
107. ICLR 2024Revisit and Outstrip Entity Alignment: A Perspective of Generative Models
108. ICLR 2024Reward Model Ensembles Help Mitigate Overoptimization
109. ICLR 2024Ring-A-Bell! How Reliable are Concept Removal Methods For Diffusion Models?
110. ICLR 2024RLCD: Reinforcement Learning from Contrastive Distillation for LM Alignment
111. ICLR 2024Robust Similarity Learning with Difference Alignment Regularization
112. ICLR 2024Robustifying and Boosting Training-Free Neural Architecture Search
113. ICLR 2024Safe and Robust Watermark Injection with a Single OoD Image
114. ICLR 2024Safe RLHF: Safe Reinforcement Learning from Human Feedback
115. ICLR 2024Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions
116. ICLR 2024SALMON: Self-Alignment with Instructable Reward Models
117. ICLR 2024Scaling for Training Time and Post-hoc Out-of-distribution Detection Enhancement
118. ICLR 2024SCHEMA: State CHangEs MAtter for Procedure Planning in Instructional Videos
119. ICLR 2024Self-Alignment with Instruction Backtranslation
120. ICLR 2024Self-contradictory Hallucinations of Large Language Models: Evaluation, Detection and Mitigation
121. ICLR 2024Self-supervised Pocket Pretraining via Protein Fragment-Surroundings Alignment
122. ICLR 2024Set Learning for Accurate and Calibrated Models
123. ICLR 2024SliceGPT: Compress Large Language Models by Deleting Rows and Columns
124. ICLR 2024Sparse Spiking Neural Network: Exploiting Heterogeneity in Timescales for Pruning Recurrent SNN
125. ICLR 2024Statistical Rejection Sampling Improves Preference Optimization
126. ICLR 2024Subtractive Mixture Models via Squaring: Representation and Learning
127. ICLR 2024Sudden Drops in the Loss: Syntax Acquisition, Phase Transitions, and Simplicity Bias in MLMs
128. ICLR 2024Task structure and nonlinearity jointly determine learned representational geometry
129. ICLR 2024Teaching Language Models to Hallucinate Less with Synthetic Tasks
130. ICLR 2024The Alignment Problem from a Deep Learning Perspective
131. ICLR 2024The Devil is in the Neurons: Interpreting and Mitigating Social Biases in Language Models
132. ICLR 2024The False Promise of Imitating Proprietary Language Models
133. ICLR 2024The importance of feature preprocessing for differentially private linear optimization
134. ICLR 2024The Lipschitz-Variance-Margin Tradeoff for Enhanced Randomized Smoothing
135. ICLR 2024The Trickle-down Impact of Reward Inconsistency on RLHF
136. ICLR 2024The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning
137. ICLR 2024Tool-Augmented Reward Modeling
138. ICLR 2024Toward effective protection against diffusion-based mimicry through score distillation
139. ICLR 2024Towards Optimal Feature-Shaping Methods for Out-of-Distribution Detection
140. ICLR 2024Towards Poisoning Fair Representations
141. ICLR 2024Towards Reliable and Efficient Backdoor Trigger Inversion via Decoupling Benign Features
142. ICLR 2024Towards Understanding Sycophancy in Language Models
143. ICLR 2024Training Socially Aligned Language Models on Simulated Social Interactions
144. ICLR 2024Understanding and Mitigating the Label Noise in Pre-training on Downstream Tasks
145. ICLR 2024Understanding Convergence and Generalization in Federated Learning through Feature Learning Theory
146. ICLR 2024Understanding the Effects of RLHF on LLM Generalisation and Diversity
147. ICLR 2024Uni-RLHF: Universal Platform and Benchmark Suite for Reinforcement Learning with Diverse Human Feedback
148. ICLR 2024Unifying Feature and Cost Aggregation with Transformers for Semantic and Visual Correspondence
149. ICLR 2024Universal Jailbreak Backdoors from Poisoned Human Feedback
150. ICLR 2024Unmasking and Improving Data Credibility: A Study with Datasets for Training Harmless Language Models

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。