ICLR 2026 Oral Paper List

LLM & Agents

Agents & Embodied

• ICLR 2026 Oral: Agent Data Protocol: Unifying Datasets for Diverse, Effective Fine-tuning of LLM Agents, arXiv, OpenReview
• ICLR 2026 Oral: Huxley-Gödel Machine: Human-Level Coding Agent Development by an Approximation of the Optimal Self-Improving Machine, arXiv, OpenReview
• ICLR 2026 Oral: MomaGraph: State-Aware Unified Scene Graphs with Vision-Language Models for Embodied Task Planning, arXiv, OpenReview
• ICLR 2026 Oral: ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data, arXiv, OpenReview
• ICLR 2026 Oral: Shoot First, Ask Questions Later? Building Rational Agents that Explore and Act Like People, arXiv, OpenReview
• ICLR 2026 Oral: Speculative Actions: A Lossless Framework for Faster AI Agents, arXiv, OpenReview

Architecture & Novel Models

• ICLR 2026 Oral: Energy-Based Transformers are Scalable Learners and Thinkers, arXiv, OpenReview
• ICLR 2026 Oral: Mamba-3: Improved Sequence Modeling using State Space Principles, arXiv, OpenReview
• ICLR 2026 Oral: Mixture-of-Experts Can Surpass Dense LLMs Under Strictly Equal Resource, arXiv, OpenReview
• ICLR 2026 Oral: Optimal Sparsity of Mixture-of-Experts Language Models for Reasoning Tasks, arXiv, OpenReview
• ICLR 2026 Oral: ParaRNN: Unlocking Parallel Training of Nonlinear RNNs for Large Language Models, arXiv, OpenReview
• ICLR 2026 Oral: Coupling Experts and Routers in Mixture-of-Experts via an Auxiliary Loss, arXiv, OpenReview
• ICLR 2026 Oral: Diffusion Language Model Knows the Answer Before It Decodes, arXiv, OpenReview
• ICLR 2026 Oral: Latent Speech-Text Transformer, arXiv, OpenReview
• ICLR 2026 Oral: MrRoPE: Mixed-radix Rotary Position Embedding, arXiv, OpenReview

Reasoning & Tool Use

• ICLR 2026 Oral: In-the-Flow Agentic System Optimization for Effective Planning and Tool Use, arXiv, OpenReview
• ICLR 2026 Oral: LLMs Get Lost In Multi-Turn Conversation, arXiv, OpenReview
• ICLR 2026 Oral: Reasoning with Sampling: Your Base Model is Smarter Than You Think, arXiv, OpenReview
• ICLR 2026 Oral: To Infinity and Beyond: Tool-Use Unlocks Length Generalization in State Space Models, arXiv, OpenReview
• ICLR 2026 Oral: In-Place Test-Time Training, arXiv, OpenReview
• ICLR 2026 Oral: On the Reasoning Abilities of Masked Diffusion Language Models, arXiv, OpenReview

Pretraining & Optimization

• ICLR 2026 Oral: Pre-training under infinite compute, arXiv, OpenReview
• ICLR 2026 Oral: Common Corpus: The Largest Collection of Ethical Data for LLM Pre-Training, arXiv, OpenReview
• ICLR 2026 Oral: AutoEP: LLMs-Driven Automation of Hyperparameter Evolution for Metaheuristic Algorithms, arXiv, OpenReview
• ICLR 2026 Oral: How Learning Rate Decay Wastes Your Best Data in Curriculum-Based LLM Pretraining, arXiv, OpenReview
• ICLR 2026 Oral: RefineStat: Efficient Exploration for Probabilistic Program Synthesis, arXiv, OpenReview
• ICLR 2026 Oral: WSM: Decay-Free Learning Rate Schedule via Checkpoint Merging for LLM Pre-training, arXiv, OpenReview

Retrieval & Utility

• ICLR 2026 Oral: Q-RAG: Long Context Multi‑Step Retrieval via Value‑Based Embedder Training, arXiv, OpenReview
• ICLR 2026 Oral: LLM DNA: Tracing Model Evolution via Functional Representations, arXiv, OpenReview
• ICLR 2026 Oral: Revela: Dense Retriever Learning via Language Modeling, arXiv, OpenReview

Post-Training & RL

• ICLR 2026 Oral: AgentGym-RL: An Open-Source Framework to Train LLM Agents for Long-Horizon Decision Making via Multi-Turn RL, arXiv, OpenReview
• ICLR 2026 Oral: Overthinking Reduction with Decoupled Rewards and Curriculum Data Scheduling, arXiv, OpenReview
• ICLR 2026 Oral: Reducing Belief Deviation in Reinforcement Learning for Active Reasoning of LLM Agents, arXiv, OpenReview
• ICLR 2026 Oral: The Coverage Principle: How Pre-Training Enables Post-Training, arXiv, OpenReview
• ICLR 2026 Oral: RAIN-Merging: A Gradient-Free Method to Enhance Instruction Following in Large Reasoning Models with Preserved Thinking Format, arXiv, OpenReview
• ICLR 2026 Oral: GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning, arXiv, OpenReview
• ICLR 2026 Oral: OpenThoughts: Data Recipes for Reasoning Models, arXiv, OpenReview
• ICLR 2026 Oral: The Art of Scaling Reinforcement Learning Compute for LLMs, arXiv, OpenReview
• ICLR 2026 Oral: TROLL: Trust Regions Improve Reinforcement Learning for Large Language Models, arXiv, OpenReview
• ICLR 2026 Oral: P-GenRM: Personalized Generative Reward Model with Test-time User-based Scaling, arXiv, OpenReview

Long Context & Decoding

• ICLR 2026 Oral: LongWriter-Zero: Mastering Ultra-Long Text Generation via Reinforcement Learning, arXiv, OpenReview
• ICLR 2026 Oral: LoongRL: Reinforcement Learning for Advanced Reasoning over Long Contexts, arXiv, OpenReview
• ICLR 2026 Oral: MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agent, arXiv, OpenReview
• ICLR 2026 Oral: Global Resolution: Optimal Multi-Draft Speculative Sampling via Convex Optimization, arXiv, OpenReview
• ICLR 2026 Oral: Intrinsic Entropy of Context Length Scaling in LLMs, arXiv, OpenReview
• ICLR 2026 Oral: Overcoming Joint Intractability with Lossless Hierarchical Speculative Decoding, arXiv, OpenReview
• ICLR 2026 Oral: ThinKV: Thought-Adaptive KV Cache Compression for Efficient Reasoning Models, arXiv, OpenReview
• ICLR 2026 Oral: $p\text{-less}$ Sampling: A Robust Hyperparameter-Free Approach for LLM Decoding, arXiv, OpenReview

Evaluation & Benchmark

• ICLR 2026 Oral: Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments, arXiv, OpenReview
• ICLR 2026 Oral: EditBench: Evaluating LLM Abilities to Perform Real-World Instructed Code Edits, arXiv, OpenReview
• ICLR 2026 Oral: OpenApps: Simulating Environment Variations to Measure UI Agent Reliability, arXiv, OpenReview
• ICLR 2026 Oral: SimuHome: A Temporal- and Environment-Aware Benchmark for Smart Home LLM Agents, arXiv, OpenReview
• ICLR 2026 Oral: SWINGARENA: Adversarial Programming Arena for Long-context GitHub Issue Solving, arXiv, OpenReview
• ICLR 2026 Oral: BIRD-INTERACT: Re-imagining Text-to-SQL Evaluation via Lens of Dynamic Interactions, arXiv, OpenReview
• ICLR 2026 Oral: AstaBench: Rigorous Benchmarking of AI Agents with a Scientific Research Suite, arXiv, OpenReview
• ICLR 2026 Oral: Train-before-Test Harmonizes Language Model Rankings, arXiv, OpenReview
• ICLR 2026 Oral: How Reliable is Language Model Micro-Benchmarking?, arXiv, OpenReview
• ICLR 2026 Oral: WebDevJudge: Evaluating (M)LLMs as Critiques for Web Development Quality, arXiv, OpenReview

Vision & Multimodal

3D & Geometry

• ICLR 2026 Oral: Depth Anything 3: Recovering the Visual Space from Any Views, arXiv, OpenReview
• ICLR 2026 Oral: DepthLM: Metric Depth from Vision Language Models, arXiv, OpenReview
• ICLR 2026 Oral: Monocular Normal Estimation via Shading Sequence Estimation, arXiv, OpenReview
• ICLR 2026 Oral: On the Generalization Capacities of MLLMs for Spatial Intelligence, arXiv, OpenReview
• ICLR 2026 Oral: Radiometrically Consistent Gaussian Surfels for Inverse Rendering, arXiv, OpenReview
• ICLR 2026 Oral: Text-to-3D by Stitching a Multi-view Reconstruction Network to a Video Generator, arXiv, OpenReview
• ICLR 2026 Oral: True Self-Supervised Novel View Synthesis is Transferable, arXiv, OpenReview
• ICLR 2026 Oral: Vid-LLM: A Compact Video-based 3D Multimodal LLM with Reconstruction–Reasoning Synergy, arXiv, OpenReview

VLM & MLLM

• ICLR 2026 Oral: Generative Universal Verifier as Multimodal Meta-Reasoner, arXiv, OpenReview
• ICLR 2026 Oral: Visual Planning: Let’s Think Only with Images, arXiv, OpenReview
• ICLR 2026 Oral: FRABench and UFEval: Unified Fine-grained Evaluation with Task and Aspect Generalization, arXiv, OpenReview
• ICLR 2026 Oral: Learning to See Before Seeing: Demystifying LLM Visual Priors from Language Pre-training, arXiv, OpenReview
• ICLR 2026 Oral: MC-Search: Evaluating and Enhancing Multimodal Agentic Search with Structured Long Reasoning Chains, arXiv, OpenReview
• ICLR 2026 Oral: MetaEmbed: Scaling Multimodal Retrieval at Test-Time with Flexible Late Interaction, arXiv, OpenReview
• ICLR 2026 Oral: Multimodal Aligned Semantic Knowledge for Unpaired Image-text Matching, arXiv, OpenReview
• ICLR 2026 Oral: Through the Lens of Contrast: Self-Improving Visual Reasoning in VLMs, arXiv, OpenReview
• ICLR 2026 Oral: Visual symbolic mechanisms: Emergent symbol processing in Vision Language Models, arXiv, OpenReview

Audio & Speech

• ICLR 2026 Oral: VibeVoice: Expressive Podcast Generation with Next-Token Diffusion, arXiv, OpenReview
• ICLR 2026 Oral: UALM: Unified Audio Language Model for Understanding, Generation and Reasoning, arXiv, OpenReview
• ICLR 2026 Oral: EmotionThinker: Prosody-Aware Reinforcement Learning for Explainable Speech Emotion Reasoning, arXiv, OpenReview
• ICLR 2026 Oral: TTSDS2: Resources and Benchmark for Evaluating Human-Quality Text to Speech Systems, arXiv, OpenReview
• ICLR 2026 Oral: WAVE: Learning Unified & Versatile Audio-Visual Embeddings with Multimodal LLM, arXiv, OpenReview
• ICLR 2026 Oral: Latent Fourier Transform, arXiv, OpenReview

Low-Level Vision & Image

• ICLR 2026 Oral: WAFT: Warping-Alone Field Transforms for Optical Flow, arXiv, OpenReview
• ICLR 2026 Oral: AnyUp: Universal Feature Upsampling, arXiv, OpenReview
• ICLR 2026 Oral: Instilling an Active Mind in Avatars via Cognitive Simulation, arXiv, OpenReview
• ICLR 2026 Oral: Reasoning as Representation: Rethinking Visual Reinforcement Learning in Image Quality Assessment, arXiv, OpenReview
• ICLR 2026 Oral: TRACE: Your Diffusion Model is Secretly an Instance Edge Detector, arXiv, OpenReview

Video

• ICLR 2026 Oral: $PhyWorldBench$: A Comprehensive Evaluation of Physical Realism in Text-to-Video Models, arXiv, OpenReview
• ICLR 2026 Oral: FlashVID: Efficient Video Large Language Models via Training-free Tree-based Spatiotemporal Token Merging, arXiv, OpenReview
• ICLR 2026 Oral: InfoTok: Adaptive Discrete Video Tokenizer via Information-Theoretic Compression, arXiv, OpenReview

RL & Decision Making

World Models & Planning

• ICLR 2026 Oral: World-In-World: World Models in a Closed-Loop World, arXiv, OpenReview
• ICLR 2026 Oral: Compositional Diffusion with Guided search for Long-Horizon Planning, arXiv, OpenReview
• ICLR 2026 Oral: TD-JEPA: Latent-predictive Representations for Zero-Shot Reinforcement Learning, arXiv, OpenReview
• ICLR 2026 Oral: Exploratory Diffusion Model for Unsupervised Reinforcement Learning, arXiv, OpenReview
• ICLR 2026 Oral: Latent Particle World Models: Self-supervised Object-centric Stochastic Dynamics Modeling, arXiv, OpenReview
• ICLR 2026 Oral: Optimistic Task Inference for Behavior Foundation Models, arXiv, OpenReview

Robot & Control

• ICLR 2026 Oral: Differentiable Model Predictive Control on the GPU, arXiv, OpenReview
• ICLR 2026 Oral: Mean Flow Policy with Instantaneous Velocity Constraint for One-step Action Generation, arXiv, OpenReview
• ICLR 2026 Oral: Rodrigues Network for Learning Robot Actions, arXiv, OpenReview

Optimization Applications

• ICLR 2026 Oral: Task-free Adaptive Meta Black-box Optimization, arXiv, OpenReview
• ICLR 2026 Oral: Discount Model Search for Quality Diversity Optimization in High-Dimensional Measure Spaces, arXiv, OpenReview
• ICLR 2026 Oral: Learning to Segment for Vehicle Routing Problems, arXiv, OpenReview
• ICLR 2026 Oral: Enhancing Generative Auto-bidding with Offline Reward Evaluation and Policy Search, arXiv, OpenReview
• ICLR 2026 Oral: Triple-BERT: Do We Really Need MARL for Order Dispatch on Ride-Sharing Platforms?, arXiv, OpenReview

Diffusion & Generative

Training & Post-Training

• ICLR 2026 Oral: DiffusionNFT: Online Diffusion Reinforcement with Forward Process, arXiv, OpenReview
• ICLR 2026 Oral: Let Features Decide Their Own Solvers: Hybrid Feature Caching for Diffusion Transformers, arXiv, OpenReview
• ICLR 2026 Oral: Neon: Negative Extrapolation From Self-Training Improves Image Generation, arXiv, OpenReview
• ICLR 2026 Oral: Pareto-Conditioned Diffusion Models for Offline Multi-Objective Optimization, arXiv, OpenReview
• ICLR 2026 Oral: Partition Generative Modeling: Masked Modeling Without Masks, arXiv, OpenReview
• ICLR 2026 Oral: SAFETY-GUIDED FLOW (SGF): A UNIFIED FRAMEWORK FOR NEGATIVE GUIDANCE IN SAFE GENERATION, arXiv, OpenReview
• ICLR 2026 Oral: Universal Inverse Distillation for Matching Models with Real-Data Supervision (No GANs), arXiv, OpenReview
• ICLR 2026 Oral: Half-order Fine-Tuning for Diffusion Model: A Recursive Likelihood Ratio Optimizer, arXiv, OpenReview

Generation Applications

• ICLR 2026 Oral: EditVerse: Unifying Image and Video Editing and Generation with In-Context Learning, arXiv, OpenReview
• ICLR 2026 Oral: MotionStream: Real-Time Video Generation with Interactive Motion Controls, arXiv, OpenReview
• ICLR 2026 Oral: NextStep-1: Toward Autoregressive Image Generation with Continuous Tokens at Scale, arXiv, OpenReview
• ICLR 2026 Oral: FlashWorld: High-quality 3D Scene Generation within Seconds, arXiv, OpenReview
• ICLR 2026 Oral: SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer, arXiv, OpenReview
• ICLR 2026 Oral: Stable Video Infinity: Infinite-Length Video Generation with Error Recycling, arXiv, OpenReview
• ICLR 2026 Oral: Generative Human Geometry Distribution, arXiv, OpenReview
• ICLR 2026 Oral: Improving Diffusion Models for Class-imbalanced Training Data via Capacity Manipulation, arXiv, OpenReview
• ICLR 2026 Oral: Locality-aware Parallel Decoding for Efficient Autoregressive Image Generation, arXiv, OpenReview

Flow & Theory

• ICLR 2026 Oral: GLASS Flows: Efficient Inference for Reward Alignment of Flow and Diffusion Models, arXiv, OpenReview
• ICLR 2026 Oral: Planner Aware Path Learning in Diffusion Language Models Training, arXiv, OpenReview
• ICLR 2026 Oral: Quotient-Space Diffusion Models, arXiv, OpenReview
• ICLR 2026 Oral: Structured Flow Autoencoders: Learning Structured Probabilistic Representations with Flow Matching, arXiv, OpenReview
• ICLR 2026 Oral: The Spacetime of Diffusion Models: An Information Geometry Perspective, arXiv, OpenReview

Theory & Optimization

Transformer & Scaling Theory

• ICLR 2026 Oral: From Markov to Laplace: How Mamba In-Context Learns Markov Chains, arXiv, OpenReview
• ICLR 2026 Oral: Quantitative Bounds for Length Generalization in Transformers, arXiv, OpenReview
• ICLR 2026 Oral: Sequences of Logits Reveal the Low Rank Structure of Language Models, arXiv, OpenReview
• ICLR 2026 Oral: Softmax Transformers are Turing-Complete, arXiv, OpenReview
• ICLR 2026 Oral: Scaling Laws and Spectra of Shallow Neural Networks in the Feature Learning Regime, arXiv, OpenReview
• ICLR 2026 Oral: Temporal superposition and feature geometry of RNNs under memory demands, arXiv, OpenReview
• ICLR 2026 Oral: Transformers are Inherently Succinct, arXiv, OpenReview
• ICLR 2026 Oral: How Do Transformers Learn to Associate Tokens: Gradient Leading Terms Bring Mechanistic Interpretability, arXiv, OpenReview

Optimization & Training

• ICLR 2026 Oral: The Polar Express: Optimal Matrix Sign Methods and their Application to the Muon Algorithm, arXiv, OpenReview
• ICLR 2026 Oral: Cross-Domain Lossy Compression via Rate- and Classification-Constrained Optimal Transport, arXiv, OpenReview
• ICLR 2026 Oral: Hyperparameter Trajectory Inference with Conditional Lagrangian Optimal Transport, arXiv, OpenReview
• ICLR 2026 Oral: Pinet: Optimizing hard-constrained neural networks with orthogonal projection layers, arXiv, OpenReview
• ICLR 2026 Oral: Non-Convex Federated Optimization under Cost-Aware Client Selection, arXiv, OpenReview
• ICLR 2026 Oral: Overparametrization bends the landscape: BBP transitions at initialization in simple Neural Networks, arXiv, OpenReview

Geometry & Dynamics

• ICLR 2026 Oral: Fast Escape, Slow Convergence: Learning Dynamics of Phase Retrieval under Power-Law Data, arXiv, OpenReview
• ICLR 2026 Oral: Navigating the Latent Space Dynamics of Neural Models, arXiv, OpenReview
• ICLR 2026 Oral: On the Wasserstein Geodesic Principal Component Analysis of probability measures, arXiv, OpenReview
• ICLR 2026 Oral: Characterizing the Discrete Geometry of ReLU Networks, arXiv, OpenReview
• ICLR 2026 Oral: Information Shapes Koopman Representation, arXiv, OpenReview

Causal & Statistical

• ICLR 2026 Oral: A Representer Theorem for Hawkes Processes via Penalized Least Squares Minimization, arXiv, OpenReview
• ICLR 2026 Oral: High-dimensional Analysis of Synthetic Data Selection, arXiv, OpenReview
• ICLR 2026 Oral: Conformal Robustness Control: A New Strategy for Robust Decision, arXiv, OpenReview
• ICLR 2026 Oral: Difficult Examples Hurt Unsupervised Contrastive Learning: A Theoretical Perspective, arXiv, OpenReview
• ICLR 2026 Oral: Causal Structure Learning in Hawkes Processes with Complex Latent Confounder Networks, arXiv, OpenReview
• ICLR 2026 Oral: Distributional Equivalence in Linear Non-Gaussian Latent-Variable Cyclic Causal Models: Characterization and Learning, arXiv, OpenReview
• ICLR 2026 Oral: InfoNCE Induces Gaussian Distribution, arXiv, OpenReview
• ICLR 2026 Oral: Non-Asymptotic Analysis of (Sticky) Track-and-Stop, arXiv, OpenReview
• ICLR 2026 Oral: Online Learning and Equilibrium Computation with Ranking Feedback, arXiv, OpenReview

Formal Reasoning & Neurosymbolic

• ICLR 2026 Oral: HATSolver: Learning Gröbner Bases with Hierarchical Attention Transformers, arXiv, OpenReview
• ICLR 2026 Oral: Premise Selection for a Lean Hammer, arXiv, OpenReview

AI4Science

Biomolecules & Chemistry

• ICLR 2026 Oral: Scaling Atomistic Protein Binder Design with Generative Pretraining and Test-Time Compute, arXiv, OpenReview
• ICLR 2026 Oral: DCFold: Efficient Protein Structure Generation with Single Forward Pass, arXiv, OpenReview
• ICLR 2026 Oral: Exploring Synthesizable Chemical Space with Iterative Pathway Refinements, arXiv, OpenReview
• ICLR 2026 Oral: mCLM: A Modular Chemical Language Model that Generates Functional and Makeable Molecules, arXiv, OpenReview
• ICLR 2026 Oral: FALCON: Few-step Accurate Likelihoods for Continuous Flows, arXiv, OpenReview

Medical & Bio Signals

• ICLR 2026 Oral: A Scalable Distributed Framework for Multimodal GigaVoxel Image Registration, arXiv, OpenReview
• ICLR 2026 Oral: MedAgentGym: A Scalable Agentic Training Environment for Code-Centric Reasoning in Biomedical Data Science, arXiv, OpenReview
• ICLR 2026 Oral: Decentralized Attention Fails Centralized Signals: Rethinking Transformers for Medical Time Series, arXiv, OpenReview
• ICLR 2026 Oral: Extending Sequence Length is Not All You Need: Effective Integration of Multimodal Signals for Gene Expression Prediction, arXiv, OpenReview
• ICLR 2026 Oral: BioX-Bridge: Model Bridging for Unsupervised Cross-Modal Knowledge Transfer across Biosignals, arXiv, OpenReview

Physics & Causal

• ICLR 2026 Oral: Fast training of accurate physics-informed neural networks without gradient descent, arXiv, OpenReview
• ICLR 2026 Oral: RealPDEBench: A Benchmark for Complex Physical Systems with Real-World Data, arXiv, OpenReview
• ICLR 2026 Oral: Exploratory Causal Inference in SAEnce, arXiv, OpenReview

Brain & Cognition

• ICLR 2026 Oral: From movement to cognitive maps: recurrent neural networks reveal how locomotor development shapes hippocampal spatial coding, arXiv, OpenReview
• ICLR 2026 Oral: Generating metamers of human scene understanding, arXiv, OpenReview
• ICLR 2026 Oral: Seeing Through the Brain: New Insights from Decoding Visual Stimuli with fMRI, arXiv, OpenReview

Safety & Alignment

Alignment & RLHF

• ICLR 2026 Oral: Multiplayer Nash Preference Optimization, arXiv, OpenReview, Note
• ICLR 2026 Oral: SafeDPO: A Simple Approach to Direct Preference Optimization with Enhanced Safety, arXiv, OpenReview, Note
• ICLR 2026 Oral: Why DPO is a Misspecified Estimator and How to Fix It, arXiv, OpenReview, Note
• ICLR 2026 Oral: Semi-Supervised Preference Optimization with Limited Feedback, arXiv, OpenReview, Note
• ICLR 2026 Oral: Token-Importance Guided Direct Preference Optimization, arXiv, OpenReview, Note
• ICLR 2026 Oral: What’s In My Human Feedback? Learning Interpretable Descriptions of Preference Data, arXiv, OpenReview, Note

Interpretability & Monitoring

• ICLR 2026 Oral: Is it Thinking or Cheating? Detecting Implicit Reward Hacking by Measuring Reasoning Effort, arXiv, OpenReview, Note
• ICLR 2026 Oral: Verifying Chain-of-Thought Reasoning via Its Computational Graph, arXiv, OpenReview, Note
• ICLR 2026 Oral: Beyond Prompt-Induced Lies: Investigating LLM Deception on Benign Prompts, arXiv, OpenReview, Note
• ICLR 2026 Oral: Reliable Weak-to-Strong Monitoring of LLM Agents, arXiv, OpenReview, Note
• ICLR 2026 Oral: Hallucination Begins Where Saliency Drops, arXiv, OpenReview, Note
• ICLR 2026 Oral: Temporal Sparse Autoencoders: Leveraging the Sequential Nature of Language for Interpretability, arXiv, OpenReview, Note
• ICLR 2026 Oral: Addressing divergent representations from causal interventions on neural networks, arXiv, OpenReview, Note
• ICLR 2026 Oral: The Shape of Adversarial Influence: Characterizing LLM Latent Spaces with Persistent Homology, arXiv, OpenReview, Note

Attack & Robustness

• ICLR 2026 Oral: CyberGym: Evaluating AI Agents’ Real-World Cybersecurity Capabilities at Scale, arXiv, OpenReview, Note
• ICLR 2026 Oral: Invisible Safety Threat: Malicious Finetuning for LLM via Steganography, arXiv, OpenReview, Note
• ICLR 2026 Oral: RedTeamCUA: Realistic Adversarial Testing of Computer-Use Agents in Hybrid Web-OS Environments, arXiv, OpenReview, Note
• ICLR 2026 Oral: Watch your steps: Dormant Adversarial Behaviors that Activate upon LLM Finetuning, arXiv, OpenReview, Note
• ICLR 2026 Oral: Veritas: Generalizable Deepfake Detection via Pattern-Aware Reasoning, arXiv, OpenReview, Note
• ICLR 2026 Oral: Steering the Herd: A Framework for LLM-based Control of Social Learning, arXiv, OpenReview, Note

Privacy & Memorization

• ICLR 2026 Oral: Spherical Watermark: Encryption-Free, Lossless Watermarking for Diffusion Models, OpenReview, Note
• ICLR 2026 Oral: Benchmarking Empirical Privacy Protection for Adaptations of Large Language Models, OpenReview, Note
• ICLR 2026 Oral: Hubble: a Model Suite to Advance the Study of LLM Memorization, arXiv, OpenReview, Note
• ICLR 2026 Oral: Differentially Private Domain Discovery, arXiv, OpenReview, Note
• ICLR 2026 Oral: Every Language Model Has a Forgery-Resistant Signature, arXiv, OpenReview, Note
• ICLR 2026 Oral: Gaussian certified unlearning in high dimensions: A hypothesis testing approach, arXiv, OpenReview, Note
• ICLR 2026 Oral: LLM Fingerprinting via Semantically Conditioned Watermarks, arXiv, OpenReview, Note
• ICLR 2026 Oral: PateGAIL++: Utility Optimized Private Trajectory Generation with Imitation Learning, OpenReview, Note

Evaluation & Benchmark

• ICLR 2026 Oral: Omni-Reward: Towards Generalist Omni-Modal Reward Modeling with Free-Form Preferences, arXiv, OpenReview, Note
• ICLR 2026 Oral: CounselBench: A Large-Scale Expert Evaluation and Adversarial Benchmarking of Large Language Models in Mental Health Question Answering, arXiv, OpenReview, Note
• ICLR 2026 Oral: AdAEM: An Adaptively and Automated Extensible Measurement of LLMs’ Value Difference, arXiv, OpenReview, Note
• ICLR 2026 Oral: EigenBench: A Comparative Behavioral Measure of Value Alignment, arXiv, OpenReview, Note

Graph & Structured

Graph Learning

• ICLR 2026 Oral: Actions Speak Louder than Prompts: A Large-Scale Study of LLMs for Graph Inference, arXiv, OpenReview
• ICLR 2026 Oral: Compactness and Consistency: A Conjoint Framework for Deep Graph Clustering, arXiv, OpenReview
• ICLR 2026 Oral: Exchangeability of GNN Representations with Applications to Graph Retrieval, arXiv, OpenReview
• ICLR 2026 Oral: Learning with Dual-level Noisy Correspondence for Multi-modal Entity Alignment, arXiv, OpenReview
• ICLR 2026 Oral: Modality-free Graph In-context Alignment, arXiv, OpenReview
• ICLR 2026 Oral: Multi-Domain Riemannian Graph Gluing for Building Graph Foundation Models, arXiv, OpenReview
• ICLR 2026 Oral: One for Two: A Unified Framework for Imbalanced Graph Classification via Dynamic Balanced Prototype, arXiv, OpenReview

Structured Data

• ICLR 2026 Oral: CauKer: Classification Time Series Foundation Models Can Be Pretrained on Synthetic Data, arXiv, OpenReview
• ICLR 2026 Oral: TabStruct: Measuring Structural Fidelity of Tabular Data, arXiv, OpenReview
• ICLR 2026 Oral: Uncover Underlying Correspondence for Robust Multi-view Clustering, arXiv, OpenReview

Other

Systems & Efficiency

• ICLR 2026 Oral: TileLang: Bridge Programmability and Performance in Modern Neural Kernels, arXiv, OpenReview
• ICLR 2026 Oral: Why Low-Precision Transformer Training Fails: An Analysis on Flash Attention, arXiv, OpenReview
• ICLR 2026 Oral: Efficient Resource-Constrained Training of Transformers via Subspace Optimization, arXiv, OpenReview
• ICLR 2026 Oral: Taming Momentum: Rethinking Optimizer States Through Low-Rank Approximation, arXiv, OpenReview
• ICLR 2026 Oral: DTO-KD: Dynamic Trade-off Optimization for Effective Knowledge Distillation, arXiv, OpenReview
• ICLR 2026 Oral: It’s All Just Vectorization: einx, a Universal Notation for Tensor Operations, arXiv, OpenReview
• ICLR 2026 Oral: Mastering Sparse CUDA Generation through Pretrained Models and Deep Reinforcement Learning, arXiv, OpenReview

Learning Methods

• ICLR 2026 Oral: Plug-and-Play Compositionality for Boosting Continual Learning with Foundation Models, arXiv, OpenReview
• ICLR 2026 Oral: On The Surprising Effectiveness of a Single Global Merging in Decentralized Learning, arXiv, OpenReview
• ICLR 2026 Oral: FIRE: Frobenius-Isometry Reinitialization for Balancing the Stability–Plasticity Tradeoff, arXiv, OpenReview
• ICLR 2026 Oral: Probabilistic Kernel Function for Fast Angle Testing, arXiv, OpenReview