Alignment and Safety papers, 2023–2026

Key findings

• Alignment and Safety went from 438 papers in 2023 to 1,882 in 2026, roughly 4.3x in three years.
• NeurIPS publishes the most Alignment and Safety work: 1,421 papers from 2023 to 2026, 31% of the four-venue total.
• The fastest-growing overlap is with Social aspects · safety: papers tagged with both rose from 0 in 2023 to 137 in 2026.
• Among the 5 topics under Alignment and Responsible AI, it ranks #1 by 2026 paper count; the fastest grower in the group, Social aspects · safety, is up 93x versus 4.3x here.

Representative papers

Spread across years and venues so every edition shows up

• 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

All Alignment and Safety papers

473 papers, page 6 of 10

251. ICLR 2025LLaMaFlex: Many-in-one LLMs via Generalized Pruning and Weight Sharing
252. ICLR 2025LLaVA-MoD: Making LLaVA Tiny via MoE-Knowledge Distillation
253. ICLR 2025LLM Unlearning via Loss Adjustment with Only Forget Data
254. ICLR 2025Local Patterns Generalize Better for Novel Anomalies
255. ICLR 2025Locality Alignment Improves Vision-Language Models
256. ICLR 2025Logicbreaks: A Framework for Understanding Subversion of Rule-based Inference
257. ICLR 2025Long-tailed Adversarial Training with Self-Distillation
258. ICLR 2025LongPO: Long Context Self-Evolution of Large Language Models through Short-to-Long Preference Optimization
259. ICLR 2025Looking Inward: Language Models Can Learn About Themselves by Introspection
260. ICLR 2025MA-RLHF: Reinforcement Learning from Human Feedback with Macro Actions
261. ICLR 2025Machine Unlearning Fails to Remove Data Poisoning Attacks
262. ICLR 2025MACPO: Weak-to-Strong Alignment via Multi-Agent Contrastive Preference Optimization
263. ICLR 2025Magnetic Preference Optimization: Achieving Last-iterate Convergence for Language Model Alignment
264. ICLR 2025Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing
265. ICLR 2025MallowsPO: Fine-Tune Your LLM with Preference Dispersions
266. ICLR 2025MAP: Multi-Human-Value Alignment Palette
267. ICLR 2025Mask-DPO: Generalizable Fine-grained Factuality Alignment of LLMs
268. ICLR 2025Maximizing the Potential of Synthetic Data: Insights from Random Matrix Theory
269. ICLR 2025Measuring memorization in RLHF for code completion
270. ICLR 2025Measuring Non-Adversarial Reproduction of Training Data in Large Language Models
271. ICLR 2025MIA-Bench: Towards Better Instruction Following Evaluation of Multimodal LLMs
272. ICLR 2025MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models
273. ICLR 2025Mind Control through Causal Inference: Predicting Clean Images from Poisoned Data
274. ICLR 2025Mitigate the Gap: Improving Cross-Modal Alignment in CLIP
275. ICLR 2025Mitigating Memorization in Language Models
276. ICLR 2025Mitigating Reward Over-Optimization in RLHF via Behavior-Supported Regularization
277. ICLR 2025Mitigating Spurious Correlations in Zero-Shot Multimodal Models
278. ICLR 2025Mitigating the Backdoor Effect for Multi-Task Model Merging via Safety-Aware Subspace
279. ICLR 2025Mix-CPT: A Domain Adaptation Framework via Decoupling Knowledge Learning and Format Alignment
280. ICLR 2025Mix-LN: Unleashing the Power of Deeper Layers by Combining Pre-LN and Post-LN
281. ICLR 2025MMDT: Decoding the Trustworthiness and Safety of Multimodal Foundation Models
282. ICLR 2025Model Editing as a Robust and Denoised variant of DPO: A Case Study on Toxicity
283. ICLR 2025Modeling Future Conversation Turns to Teach LLMs to Ask Clarifying Questions
284. ICLR 2025Monitoring Latent World States in Language Models with Propositional Probes
285. ICLR 2025Montessori-Instruct: Generate Influential Training Data Tailored for Student Learning
286. ICLR 2025Moral Alignment for LLM Agents
287. ICLR 2025More RLHF, More Trust? On The Impact of Preference Alignment On Trustworthiness
288. ICLR 2025Multi-objective antibody design with constrained preference optimization
289. ICLR 2025Navigating Neural Space: Revisiting Concept Activation Vectors to Overcome Directional Divergence
290. ICLR 2025Near-Exact Privacy Amplification for Matrix Mechanisms
291. ICLR 2025Neural Interactive Proofs
292. ICLR 2025NNsight and NDIF: Democratizing Access to Open-Weight Foundation Model Internals
293. ICLR 2025No Preference Left Behind: Group Distributional Preference Optimization
294. ICLR 2025OASIS Uncovers: High-Quality T2I Models, Same Old Stereotypes
295. ICLR 2025On a Connection Between Imitation Learning and RLHF
296. ICLR 2025On Calibration of LLM-based Guard Models for Reliable Content Moderation
297. ICLR 2025On Evaluating the Durability of Safeguards for Open-Weight LLMs
298. ICLR 2025On Large Language Model Continual Unlearning
299. ICLR 2025On Linear Representations and Pretraining Data Frequency in Language Models
300. ICLR 2025On Targeted Manipulation and Deception when Optimizing LLMs for User Feedback

FAQ

How many Alignment and Safety papers were published at NeurIPS, ICLR, ICML and CVPR from 2023 to 2026?

4,558 in total: 438 in 2023, 875 in 2024, 1,363 in 2025, 1,882 in 2026.

How many Alignment and Safety papers are at NeurIPS 2026?

NeurIPS 2026 has 480 Alignment and Safety papers, versus 455 the year before.

How many Alignment and Safety papers are at ICLR 2026?

ICLR 2026 has 623 Alignment and Safety papers, versus 473 the year before.

How many Alignment and Safety papers are at ICML 2026?

ICML 2026 has 551 Alignment and Safety papers, versus 300 the year before.

How many Alignment and Safety papers are at CVPR 2026?

CVPR 2026 has 228 Alignment and Safety papers, versus 135 the year before.

Which conference publishes the most Alignment and Safety papers?

NeurIPS, with 1,421 papers from 2023 to 2026.