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 551 篇，第 10 / 12 页

451. ICML 2026State-Dependent Safety Failures in Multi-Turn Language Model Interaction
452. ICML 2026Steal the Patch Size: Adversarially Manipulate Vision Language Models
453. ICML 2026Steering at the Source: Style Modulation Heads for Robust Persona Control
454. ICML 2026Steering Beyond the Support: Adversarial Training on Unsupervised Jailbroken Activation Simulation
455. ICML 2026Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning
456. ICML 2026SteeringSafety: Benchmarking Representation Steering in LLMs Across Safety Perspectives
457. ICML 2026STLA: Spatiotemporal Lookahead Alignment for Post-Training Quantization
458. ICML 2026Stochastic Minimum-Cost Reach-Avoid Reinforcement Learning
459. ICML 2026Structured Multi-step Jailbreaking under a Hamiltonian Generative Formulation
460. ICML 2026Submodular Optimization for Minimal Augmentation in Robust Language Model Alignment
461. ICML 2026Supervised Classification Heads as Semantic Prototypes: Unlocking Vision-Language Alignment via Weight Recycling
462. ICML 2026Sycophancy Towards Researchers Drives Performative Misalignment
463. ICML 2026T-GINEE: A Tensor-Based Multi-Graph Representation Learning
464. ICML 2026T-POP: Test-Time Personalization with Online Preference Feedback
465. ICML 2026Tackling Fake Forgetting through Uncertainty Quantification
466. ICML 2026Tackling Length Inflation Without Trade-offs: Group Relative Reward Rescaling for Reinforcement Learning
467. ICML 2026Talk, Judge, Cooperate: Gossip-Driven Indirect Reciprocity in Self-Interested LLM Agents
468. ICML 2026Task-Aware Preference Calibration for Direct Preference Optimization
469. ICML 2026Task-Awareness Improves LLM Generations and Uncertainty
470. ICML 2026Temporal Preference Optimization for Unsupervised Retrieval
471. ICML 2026Temporal Self-Rewarding Language Models: Decoupling Chosen-Rejected via Past-Future
472. ICML 2026Test-Time Detoxification without Training or Learning Anything
473. ICML 2026The Assistant Axis: Situating and Stabilizing the Default Persona of Language Models
474. ICML 2026The Axiomatic Value of Regularization in AI Alignment from Human Preferences
475. ICML 2026The benefits of full data shuffle, now with optimal I/O cost: $k$-wise independence and matrix transposition to the rescue
476. ICML 2026The Differences Between Direct Alignment Algorithms are a Blur
477. ICML 2026The Geometric Mechanics of Contrastive Representation Learning: Alignment Potentials, Entropic Dispersion, and Cross-Modal Divergence
478. ICML 2026The Geometry of Projection Heads: Conditioning, Invariance, and Collapse
479. ICML 2026The Geometry of Updates: Fisher Alignment at Vocabulary Scale
480. ICML 2026The Heterogeneous Safety Impacts of Benign Multilingual Fine-Tuning
481. ICML 2026The Hidden Link between RLHF and Contrastive Learning
482. ICML 2026The Implicit Bias of Adam and Muon on Smooth Homogeneous Neural Networks
483. ICML 2026The Labyrinth and the Thread: Rethinking Regularizations in Sequential Knowledge Editing for Large Language Models
484. ICML 2026The Oversight Game: Learning to Cooperatively Balance an AI Agent's Safety and Autonomy
485. ICML 2026The Pareto-optimal Trade-off between Regret and Statistical Inference in Linear Stochastic Bandits under Safety Constraints
486. ICML 2026The Perception–Physics Paradox: Probing Scientific Alignment with TC-Bench
487. ICML 2026The Personality Illusion: Revealing Dissociation Between Self-Reports & Behavior in LLMs
488. ICML 2026The Realignment Problem: When Right becomes Wrong in LLMs
489. ICML 2026The Sign Estimator: Preference Modeling for LLM Alignment under Heterogeneity
490. ICML 2026The Theory and Practice of MAP Inference over Non-Convex Constraints
491. ICML 2026The Trojan Knowledge: Bypassing Commercial LLM Guardrails via Harmless Prompt Weaving and Adaptive Tree Search
492. ICML 2026Theoretical Analysis of Sparse Optimization with Reparameterization, Weight Decay, and Adaptive Learning Rate
493. ICML 2026Theoretical Challenges in Learning for Branch-and-Cut
494. ICML 2026Think Twice Before You Act: Enhancing Agent Behavioral Safety with Thought Correction
495. ICML 2026Threshold-Guided Optimization for Visual Generative Models
496. ICML 2026Through the Stealth Lens: Attention-Aware Defenses Against Poisoning in RAG
497. ICML 2026Time Series, Vision, and Language: Exploring the Limits of Alignment in Contrastive Representation Spaces
498. ICML 2026TMD-Bench: A Multi-Level Evaluation Paradigm for Music–Dance Co-Generation
499. ICML 2026TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching
500. ICML 2026Toward Safe Quantization-Aware Fine-tuning: Understanding and Mitigating Safety Alignment Degradation

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。