对齐与安全论文趋势 2023–2026
RLHF、偏好优化、红队与安全对齐
统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 4,558 篇被归入该主题的论文。
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。
各会议论文数
按年份和会议统计
| 年份 | NeurIPS | ICLR | ICML | CVPR | 合计 |
|---|---|---|---|---|---|
| 2023 | 138 | 87 | 101 | 112 | 438 |
| 2024 | 348 | 156 | 246 | 125 | 875 |
| 2025 | 455 | 473 | 300 | 135 | 1,363 |
| 2026 | 480 | 623 | 551 | 228 | 1,882 |
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 1,421 篇，第 5 / 29 页
- NeurIPS 2026Inference Time Nash Alignment
- NeurIPS 2026Inference-time Alignment via Sparse Junction Steering
- NeurIPS 2026Inference-Time Vulnerability Beyond Shallow Safety: Alignment Along Generation Trajectories
- NeurIPS 2026Instance-Dependent Bandit Convex Optimization in One Dimension
- NeurIPS 2026Internal Safety Collapse in Frontier Large Language Models
- NeurIPS 2026Interpretable Multiway-Split Trees
- NeurIPS 2026Intersectional Fairness via Mixed-Integer Optimization
- NeurIPS 2026Introspective Coupling: LMs Learn to Explain Themselves Better Than Their Training Targets
- NeurIPS 2026Invariance and Body-Order Compose Additively: Minimax Rates on $\mathrm{SO}(3)^n$
- NeurIPS 2026ITO: Multi-View Alignment and Training-Time Fusion for Image-Text Pretraining
- NeurIPS 2026JailBound: A FOL-Guided Jailbreak Evaluation Framework for Revealing Safety Boundaries of LLMs
- NeurIPS 2026Jailbreak susceptibility prediction and mitigation via the behavioral geometry of models
- NeurIPS 2026Jointly Robust Fairness: Overcoming Simultaneous Label and Attribute Noise
- NeurIPS 2026JudgmentBench: Comparing Rubric and Preference Evaluation for Quality Assessment
- NeurIPS 2026L-FAME: Longitudinal Focused Attention Meditation EEG Dataset and Benchmark
- NeurIPS 2026Latency-Conditioned Selection Bias in RLHF
- NeurIPS 2026Leaderboard Hacking: Preference-Based Model Evaluations are Vulnerable to Manipulation
- NeurIPS 2026Learning Data-free Universal Adversarial Perturbation with Hybrid Priors and Gradient-Guided Sharpness Regularization
- NeurIPS 2026Learning How to Cube
- NeurIPS 2026Learning Rate Transfer in Normalized Transformers
- NeurIPS 2026Learning to Decide with AI Assistance under Human-Alignment
- NeurIPS 2026Leveraging Psychophysical Attentional Distribution for Gaze-Augmented Reward Modeling
- NeurIPS 2026Leviathan: Decoupling Input and Output Representations in Language Models
- NeurIPS 2026Limits and Potential of Score-Based Data Valuation: Redundancy, Complementarity, and Non-Monotonicity
- NeurIPS 2026LiSA: Lifelong Safety Adaptation via Conservative Policy Induction
- NeurIPS 2026Listening to the Wise Few: Query–Key Alignment Unlocks Latent Correct Answers in Large Language Models
- NeurIPS 2026LLM Active Alignment: A Nash Equilibrium Perspective
- NeurIPS 2026LLM Alignment--Utility Asymmetry under Semantic-Preserving Transformations
- NeurIPS 2026Local Sparsity Enables Unsupervised LLM Safety Detection
- NeurIPS 2026Localizing and Repairing Sparse-Prompt Failure in SAM Decoders via Box-to-Point Counterfactual
- NeurIPS 2026Logit-Gap Steering: A Forward-Pass Diagnostic for Alignment Robustness
- NeurIPS 2026Look Before You Leap: Self-Evolving Clinical Reasoning with Psychometric Preference Optimization for Radiology Report Generation
- NeurIPS 2026Looking Through the Mirror: Minimax-Optimal Regularized Regrets in Online Learning and Bandits
- NeurIPS 2026LoopWeaver: Weaving Feedback Loops into Hierarchical Generation under Constraints
- NeurIPS 2026LOSCAR-SGD: Local SGD with Communication-Computation Overlap and Delay-Corrected Sparse Model Averaging
- NeurIPS 2026Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment
- NeurIPS 2026LucidNFT: LR-Anchored Multi-Reward Preference Optimization for Flow-Based Real-World Super-Resolution
- NeurIPS 2026Manifold Drift in Flow Preference Optimization: A Root Cause of Reward Hacking
- NeurIPS 2026Manifold-weighted neural networks
- NeurIPS 2026MaPPO: Maximum a Posteriori Preference Optimization with Prior Knowledge
- NeurIPS 2026MAPS: Margin-Aware Priors and Verifier-Guided Search for Embodied Planning
- NeurIPS 2026Margin Dynamics for Large Language Model Alignment
- NeurIPS 2026Masked Diffusion Language Agents for Tool-Integrated Chemical Reasoning
- NeurIPS 2026Matching-while-Decoding: Enhancing Template-Free Retrosynthesis via Explicit Structural Alignment
- NeurIPS 2026MATO: Multi-objective Personalized Alignment with Test-time Optimization for Large Language Models
- NeurIPS 2026Matryoshka Transcoders and Hierarchy Misalignment: When SAE Absorption Protection Does Not Transfer
- NeurIPS 2026MedKIT: Evaluating Knowledge Integration and Generalization in Large Language Models
- NeurIPS 2026Merging RLVR-Trained Experts via Policy-Shift-Guided Spectral Alignment
- NeurIPS 2026Meta-Learning Preferences for Multilingual LLM Alignment
- NeurIPS 2026META-PAP: Meta-learning for Prompt-aware Preference Pairing in LLM Alignment
在 Claude、Cursor 里直接问这个趋势
接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。
可以直接复制这句话问 AI
用 paperlist 查对齐与安全在 NeurIPS 2023–2026 的论文趋势，哪一年涨得最快？
AI 会调用的工具：get_topic_stats {"topic": "alignment-and-safety"}
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。