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 1,363 篇，第 20 / 28 页

951. NeurIPS 2025Avoiding exp(R) scaling in RLHF through Preference-based Exploration
952. NeurIPS 2025Backdoor Cleaning without External Guidance in MLLM Fine-tuning
953. NeurIPS 2025BackdoorLLM: A Comprehensive Benchmark for Backdoor Attacks and Defenses on Large Language Models
954. NeurIPS 2025Bag of Tricks for Inference-time Computation of LLM Reasoning
955. NeurIPS 2025Balancing Positive and Negative Classification Error Rates in Positive-Unlabeled Learning
956. NeurIPS 2025Bandit and Delayed Feedback in Online Structured Prediction
957. NeurIPS 2025Best-of-N Jailbreaking
958. NeurIPS 2025Better Estimation of the Kullback--Leibler Divergence Between Language Models
959. NeurIPS 2025Beyond Expectations: Quantile-Guided Alignment for Risk-Calibrated Language Models
960. NeurIPS 2025Beyond Oracle: Verifier-Supervision for Instruction Hierarchy in Reasoning and Instruction-Tuned LLMs
961. NeurIPS 2025Beyond Scalars: Concept-Based Alignment Analysis in Vision Transformers
962. NeurIPS 2025Bi-Level Decision-Focused Causal Learning for Large-Scale Marketing Optimization: Bridging Observational and Experimental Data
963. NeurIPS 2025BLEUBERI: BLEU is a surprisingly effective reward for instruction following
964. NeurIPS 2025Boosting Adversarial Transferability with Spatial Adversarial Alignment
965. NeurIPS 2025Brain-tuning Improves Generalizability and Efficiency of Brain Alignment in Speech Models
966. NeurIPS 2025Bridging Crypto with ML-based Solvers: the SAT Formulation and Benchmarks
967. NeurIPS 2025Bridging Human and LLM Judgments: Understanding and Narrowing the Gap
968. NeurIPS 2025Bridging Theory and Practice in Link Representation with Graph Neural Networks
969. NeurIPS 2025Bridging Time and Linguistics: LLMs as Time Series Analyzer through Symbolization and Segmentation
970. NeurIPS 2025C-SafeGen: Certified Safe LLM Generation with Claim-Based Streaming Guardrails
971. NeurIPS 2025Calibrating Translation Decoding with Quality Estimation on LLMs
972. NeurIPS 2025CaliGCL: Calibrated Graph Contrastive Learning via Partitioned Similarity and Consistency Discrimination
973. NeurIPS 2025CALM: Culturally Self-Aware Language Models
974. NeurIPS 2025CAMILA: Context-Aware Masking for Image Editing with Language Alignment
975. NeurIPS 2025Can DPO Learn Diverse Human Values? A Theoretical Scaling Law
976. NeurIPS 2025CARE: Decoding-Time Safety Alignment via Rollback and Introspection Intervention
977. NeurIPS 2025CARES: Comprehensive Evaluation of Safety and Adversarial Robustness in Medical LLMs
978. NeurIPS 2025Certifying Concavity and Monotonicity in Games via Sum-of-Squares Hierarchies
979. NeurIPS 2025Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment
980. NeurIPS 2025Checklists Are Better Than Reward Models For Aligning Language Models
981. NeurIPS 2025Clean First, Align Later: Benchmarking Preference Data Cleaning for Reliable LLM Alignment
982. NeurIPS 2025Clip-and-Verify: Linear Constraint-Driven Domain Clipping for Accelerating Neural Network Verification
983. NeurIPS 2025Closed-Form Training Dynamics Reveal Learned Features and Linear Structure in Word2Vec-like Models
984. NeurIPS 2025Collective Bargaining in the Information Economy Can Address AI-Driven Power Concentration
985. NeurIPS 2025Combinatorial Ski Rental Problem: Robust and Learning-Augmented Algorithms
986. NeurIPS 2025Combining Cost Constrained Runtime Monitors for AI Safety
987. NeurIPS 2025Comparison requires valid measurement: Rethinking attack success rate comparisons in AI red teaming
988. NeurIPS 2025ComPO: Preference Alignment via Comparison Oracles
989. NeurIPS 2025Composition and Alignment of Diffusion Models using Constrained Learning
990. NeurIPS 2025Comprehensive Assessment and Analysis for NSFW Content Erasure in Text-to-Image Diffusion models
991. NeurIPS 2025Compress Large Language Models via Collaboration Between Learning and Matrix Approximation
992. NeurIPS 2025Computation and Memory-Efficient Model Compression with Gradient Reweighting
993. NeurIPS 2025Concentration and excess risk bounds for imbalanced classification with synthetic oversampling
994. NeurIPS 2025Confidence-Aware With Prototype Alignment for Partial Multi-label Learning
995. NeurIPS 2025Conformal Arbitrage: Risk-Controlled Balancing of Competing Objectives in Language Models
996. NeurIPS 2025Connecting Neural Models Latent Geometries with Relative Geodesic Representations
997. NeurIPS 2025Consistent Supervised-Unsupervised Alignment for Generalized Category Discovery
998. NeurIPS 2025Continuity and Isolation Lead to Doubts or Dilemmas in Large Language Models
999. NeurIPS 2025Contrastive Learning with Data Misalignment: Feature Purity, Training Dynamics and Theoretical Generalization Guarantees
1000. NeurIPS 2025CoP: Agentic Red-teaming for Large Language Models using Composition of Principles

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。