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 1,421 篇，第 15 / 29 页

701. NeurIPS 2025Language Models Are Capable of Metacognitive Monitoring and Control of Their Internal Activations
702. NeurIPS 2025Large language models can learn and generalize steganographic chain-of-thought under process supervision
703. NeurIPS 2025LARGO: Latent Adversarial Reflection through Gradient Optimization for Jailbreaking LLMs
704. NeurIPS 2025LASeR: Learning to Adaptively Select Reward Models with Multi-Arm Bandits
705. NeurIPS 2025Latent Harmony: Synergistic Unified UHD Image Restoration via Latent Space Regularization and Controllable Refinement
706. NeurIPS 2025LayerNavigator: Finding Promising Intervention Layers for Efficient Activation Steering in Large Language Models
707. NeurIPS 2025Learning “Partner-Aware” Collaborators in Multi-Party Collaboration
708. NeurIPS 2025Learning a Cross-Modal Schrödinger Bridge for Visual Domain Generalization
709. NeurIPS 2025Learning single index models via harmonic decomposition
710. NeurIPS 2025Learning to Add, Multiply, and Execute Algorithmic Instructions Exactly with Neural Networks
711. NeurIPS 2025Learning to Steer: Input-dependent Steering for Multimodal LLMs
712. NeurIPS 2025Less is More: Improving LLM Alignment via Preference Data Selection
713. NeurIPS 2025Let a Neural Network be Your Invariant
714. NeurIPS 2025Leveraging Importance Sampling to Detach Alignment Modules from Large Language Models
715. NeurIPS 2025Leveraging robust optimization for llm alignment under distribution shifts
716. NeurIPS 2025LeVo: High-Quality Song Generation with Multi-Preference Alignment
717. NeurIPS 2025Lifelong Safety Alignment for Language Models
718. NeurIPS 2025LLM Safety Alignment is Divergence Estimation in Disguise
719. NeurIPS 2025LLMs Encode Harmfulness and Refusal Separately
720. NeurIPS 2025LongVPO: From Anchored Cues to Self-Reasoning for Long-Form Video Preference Optimization
721. NeurIPS 2025Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy and Research
722. NeurIPS 2025macOSWorld: A Multilingual Interactive Benchmark for GUI Agents
723. NeurIPS 2025Many LLMs Are More Utilitarian Than One
724. NeurIPS 2025Matchings Under Biased and Correlated Evaluations
725. NeurIPS 2025Measuring and Guiding Monosemanticity
726. NeurIPS 2025Mechanism Design for LLM Fine-tuning with Multiple Reward Models
727. NeurIPS 2025Median Selection with Noisy and Structural Information
728. NeurIPS 2025Mesh-RFT: Enhancing Mesh Generation via Fine-grained Reinforcement Fine-Tuning
729. NeurIPS 2025Meta-Learning Objectives for Preference Optimization
730. NeurIPS 2025MGUP: A Momentum-Gradient Alignment Update Policy for Stochastic Optimization
731. NeurIPS 2025Military AI Needs Technically-Informed Regulation to Safeguard AI Research and its Applications
732. NeurIPS 2025Mitigating Hallucination Through Theory-Consistent Symmetric Multimodal Preference Optimization
733. NeurIPS 2025Mitigating Reward Over-optimization in Direct Alignment Algorithms with Importance Sampling
734. NeurIPS 2025Mitigating Sexual Content Generation via Embedding Distortion in Text-conditioned Diffusion Models
735. NeurIPS 2025Mix Data or Merge Models? Balancing the Helpfulness, Honesty, and Harmlessness of Large Language Model via Model Merging
736. NeurIPS 2025MJ-Video: Benchmarking and Rewarding Video Generation with Fine-Grained Video Preference
737. NeurIPS 2025Model Inversion with Layer-Specific Modeling and Alignment for Data-Free Continual Learning
738. NeurIPS 2025Model–Behavior Alignment under Flexible Evaluation: When the Best-Fitting Model Isn’t the Right One
739. NeurIPS 2025Multi-step Visual Reasoning with Visual Tokens Scaling and Verification
740. NeurIPS 2025Multimodal Disease Progression Modeling via Spatiotemporal Disentanglement and Multiscale Alignment
741. NeurIPS 2025Nabla-R2D3: Effective and Efficient 3D Diffusion Alignment with 2D Rewards
742. NeurIPS 2025Neural Collapse is Globally Optimal in Deep Regularized ResNets and Transformers
743. NeurIPS 2025Neural Thermodynamics: Entropic Forces in Deep and Universal Representation Learning
744. NeurIPS 2025NeuSymEA: Neuro-symbolic Entity Alignment via Variational Inference
745. NeurIPS 2025No Loss, No Gain: Gated Refinement and Adaptive Compression for Prompt Optimization
746. NeurIPS 2025Noise Hypernetworks: Amortizing Test-Time Compute in Diffusion Models
747. NeurIPS 2025Non-monotone Submodular Optimization: $p$-Matchoid Constraints and Fully Dynamic Setting
748. NeurIPS 2025NSNQuant: A Double Normalization Approach for Calibration-Free Low-Bit Vector Quantization of KV Cache
749. NeurIPS 2025nvBench 2.0: Resolving Ambiguity in Text-to-Visualization through Stepwise Reasoning
750. NeurIPS 2025Object-centric binding in Contrastive Language-Image Pretraining

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。