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 623 篇，第 4 / 13 页

151. ICLR 2026Don't Forget Its Variance! The Minimum Path Variance Principle for Accurate and Stable Score-Based Models
152. ICLR 2026Don't Shift the Trigger: Robust Gradient Ascent for Backdoor Unlearning
153. ICLR 2026Doubly-Regressing Approach for Subgroup Fairness
154. ICLR 2026Doubly-Robust LLM-as-a-Judge: Externally Valid Estimation with Imperfect Personas
155. ICLR 2026Downgrade to Upgrade: Optimizer Simplification Enhances Robustness in LLM Unlearning
156. ICLR 2026DreamCS: Geometry-Aware Text-to-3D Generation with Unpaired 3D Reward Supervision
157. ICLR 2026DRIFT: Divergent Response in Filtered Transformations for Robust Adversarial Defense
158. ICLR 2026DRIFT: Learning from Abundant User Dissatisfaction in Real-World Preference Learning
159. ICLR 2026Dropping Just a Handful of Preferences Can Change Top Large Language Model Rankings
160. ICLR 2026DrugTrail: Explainable Drug Discovery via Structured Reasoning and Druggability‑Tailored Preference Optimization
161. ICLR 2026Dual Randomized Smoothing: Beyond Global Noise Variance
162. ICLR 2026Dual-Branch Representations with Dynamic Gated Fusion and Triple-Granularity Alignment for Deep Multi-View Clustering
163. ICLR 2026Dual-IPO: Dual-Iterative Preference Optimization for Text-to-Video Generation
164. ICLR 2026Dual-Path Condition Alignment for Diffusion Transformers
165. ICLR 2026Dual-Space Smoothness for Robust and Balanced LLM Unlearning
166. ICLR 2026DUET: DISTILLED LLM UNLEARNING FROM AN EFFICIENTLY CONTEXTUALIZED TEACHER
167. ICLR 2026DuPO: Enabling Reliable Self-Verification via Dual Preference Optimization
168. ICLR 2026DVLA-RL: Dual-Level Vision–Language Alignment with Reinforcement Learning Gating for Few-Shot Learning
169. ICLR 2026DynaGuard: A Dynamic Guardian Model With User-Defined Policies
170. ICLR 2026Dynamic Reflections: Probing Video Representations with Text Alignment
171. ICLR 2026Dyslexify: A Mechanistic Defense Against Typographic Attacks in CLIP
172. ICLR 2026EAMET: ROBUST MASSIVE MODEL EDITING VIA EMBEDDING ALIGNMENT OPTIMIZATION
173. ICLR 2026Early Signs of Steganographic Capabilities in Frontier LLMs
174. ICLR 2026EasyTune: Efficient Step-Aware Fine-Tuning for Diffusion-Based Motion Generation
175. ICLR 2026EditLens: Quantifying the Extent of AI Editing in Text
176. ICLR 2026EigenBench: A Comparative Behavioral Measure of Value Alignment
177. ICLR 2026ELEPHANT: Measuring and understanding social sycophancy in LLMs
178. ICLR 2026Eliciting Harmful Capabilities by Fine-Tuning on Safeguarded Outputs
179. ICLR 2026Eliminating Inductive Bias in Reward Models with Information-Theoretic Guidance
180. ICLR 2026Emergent Misalignment is Easy, Narrow Misalignment is Hard
181. ICLR 2026EmotionHallucer: Evaluating Emotion Hallucinations in Multimodal Large Language Models
182. ICLR 2026Enabling Fine-Tuning of Direct Feedback Alignment via Feedback-Weight Matching
183. ICLR 2026Enforcing Axioms for AI Alignment under Loss-Based Rules
184. ICLR 2026Enhancing Image-Conditional Coverage in Segmentation: Adaptive Thresholding via Differentiable Miscoverage Loss
185. ICLR 2026Enhancing Language Model Reasoning with Structured Multi-Level Modeling
186. ICLR 2026Enhancing Persona Following at Decoding Time via Dynamic Importance Estimation for Role-Playing Agents
187. ICLR 2026Enhancing Trustworthiness of Fine-Tuned LLMs via Regularized Subset Selection
188. ICLR 2026Erase or Hide? Suppressing Spurious Unlearning Neurons for Robust Unlearning
189. ICLR 2026Escaping Policy Contraction: Contraction-Aware PPO (CaPPO) for Stable Language Model Fine-Tuning
190. ICLR 2026Estimating Worst-Case Frontier Risks of Open-Weight LLMs
191. ICLR 2026Evaluating and Improving Cultural Awareness of Reward Models for LLM Alignment
192. ICLR 2026EventFlash: Towards Efficient MLLMs for Event-Based Vision
193. ICLR 2026Every Language Model Has a Forgery-Resistant Signature
194. ICLR 2026Evidence for Limited Metacognition in LLMs
195. ICLR 2026Evolution and compression in LLMs: on the emergence of human-aligned categorization
196. ICLR 2026Expert Merging: Model Merging with Unsupervised Expert Alignment and Importance-Guided Layer Chunking
197. ICLR 2026Expertise Can Be Helpful for Reinforcement Learning-based Macro Placement
198. ICLR 2026ExpGuard: LLM Content Moderation in Specialized Domains
199. ICLR 2026Explainable LLM Unlearning through Reasoning
200. ICLR 2026Exploring the Basin-Like Loss Landscape in Large 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。