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 1,882 篇，第 37 / 38 页

1801. NeurIPS 2026Stealth Apart, Harm Together: Skill Cascading Attacks on Skill-Based Agent Systems
1802. NeurIPS 2026Steer2Edit: From Activation Steering to Component-Level Editing
1803. NeurIPS 2026Steering Frozen LLMs: Adaptive Social Alignment via Online Prompt Routing
1804. NeurIPS 2026StiCAS: Compositional Activation Steering via Stiefel Manifold Coordinate Transport
1805. NeurIPS 2026Stitched Value Model for Diffusion Alignment
1806. NeurIPS 2026Stochastic Optimization with Random Search
1807. NeurIPS 2026STParOpt: An End-to-End Framework for Execution-Aware Parallelism Inference and Optimized CUDA Migration
1808. NeurIPS 2026Strategic Causal Policy Learning: Welfare, Safety, and Fairness Thresholds
1809. NeurIPS 2026Strategic Feature Selection and Regularization
1810. NeurIPS 2026Stress Testing Chain-of-Thought Monitoring Against Covert Misalignment
1811. NeurIPS 2026Strong Helps Weak: Directional Cross-Modal Alignment Transfer in Multi-modal LLMs
1812. NeurIPS 2026StyleRoute: Diffusion Style Transfer via Regional Routing and Conflict-aware Projection
1813. NeurIPS 2026Subliminal Learning as Trait-Direction Drift: A Mechanism and Targeted Control under SFT Distillation
1814. NeurIPS 2026Supervised Distributional Reduction via Optimal Transport and Dependence Maximization
1815. NeurIPS 2026Surprisingly consistent failures of post-hoc OOD detectors reveal security concerns in open-set recognition with modern CNNs and ViTs, stemming from representation mismatch
1816. NeurIPS 2026SynGeo: Synergizing Seeing and Proving through Revisable Geometric States
1817. NeurIPS 2026Synthetic Anchor-Assisted Prototype Alignment for Heterogeneous Federated Learning
1818. NeurIPS 2026TailGuard: Subgroup Tail Coverage Theory for Safe LLM Alignment
1819. NeurIPS 2026Take It or Leave It: Intent-Controlled Partial Optimal Transport
1820. NeurIPS 2026Tcell: Mitigating Harmful Fine-tuning for Large Language Models via Gradient Alignment
1821. NeurIPS 2026Temporal Concentration from Rollout Errors: Implicit Preference Optimization For Text-to-Video Diffusion
1822. NeurIPS 2026Temporal Prototype Alignment for Frozen-Feature Dataset Distillation
1823. NeurIPS 2026Text as Partial Constraint: Core–Residual Alignment for Robust Vision–Language Learning
1824. NeurIPS 2026The Alignment Illusion in Multimodal Large Language Models
1825. NeurIPS 2026The Alignment Tax Concentrates in Output Projections
1826. NeurIPS 2026The Block Catastrophe of Marginal Calibration in Certified Requirements Traceability
1827. NeurIPS 2026The Constitutional Coverage Trilemma in AI Governance
1828. NeurIPS 2026The Geometry of Alignment Collapse: When Fine-Tuning Breaks Safety
1829. NeurIPS 2026The Kernel Reality Check: Benchmarking and Distilling Efficient Attention at Scale
1830. NeurIPS 2026The Many Faces of On-Policy Distillation: Pitfalls, Mechanisms, and Fixes
1831. NeurIPS 2026The Path Not Taken: RLVR Learns Off the Principals
1832. NeurIPS 2026The Piggyback Hypothesis of Generalization: Explaining and Mitigating Emergent Misalignment
1833. NeurIPS 2026The Post-Training Dilemma: Why We Should Rethink the Sequential SFT-RL Paradigm
1834. NeurIPS 2026The Structural Bias of $\ell_2$-Regularized Cross-Entropy Heads in Class-Incremental Learning: Characterization, Limits, and a Gauge-Anchoring Fix
1835. NeurIPS 2026The Value of Covariance Matching in Gaussian DDPMs and the Lanczos Sampler
1836. NeurIPS 2026TherapyGym: Evaluating and Aligning Clinical Fidelity and Safety in Therapy Chatbots
1837. NeurIPS 2026There are Levels to It: Red Teaming LLMs with Hierarchical Reinforcement Learning
1838. NeurIPS 2026ThinkSafe: Self-Generated Safety Alignment for Reasoning Models
1839. NeurIPS 2026Tight PAC-Bayes Generalisation Guarantees for Large Language Model Safety Monitoring
1840. NeurIPS 2026Too Smart to Teach: The Articulability Ceiling in Language Models
1841. NeurIPS 2026Towards Multi-Human-Value Alignment via Value Localization in LLMs
1842. NeurIPS 2026Towards Understanding the Power and Limits of the Muon Optimizer: A River-Valley Perspective
1843. NeurIPS 2026TPO: Tri-level Distributionally Robust Learning for OOD Direct Preference Optimization
1844. NeurIPS 2026Training-Free 3D Editing via Feature-Divergence Localization and Trajectory Correction
1845. NeurIPS 2026Training-Free Cultural Alignment of Large Language Models via Persona Disagreement
1846. NeurIPS 2026Tree-Ensemble Repair through Constrained Optimization
1847. NeurIPS 2026Tree-Guided Identify Then Exploit: A Unified Framework of Pure Exploration and Regret Minimization for Dueling Bandits
1848. NeurIPS 2026Tree-Structured Synergy of Large Language Models and Bayesian Optimization for Efficient CASH
1849. NeurIPS 2026TROPT: An Open Framework for Unifying and Advancing Discrete Text Optimization
1850. NeurIPS 2026Two Speeds of Learning: A Representation-Readout Decomposition of Grokking and Double Descent

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。