NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
7,093 results for “keyword:"language models"” in ICLR 2027. 1,669 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]One Pixel Wider, a Different Click: How Image Preprocessing Destabilizes GUI Grounding▼5.9k2.50
[PDF]AnchorGroup: Object-Anchored Video Understanding Benchmark with an Agentic Annotation Loop▲3.8k2.50
[PDF]Beyond Data Quantity: From Bootstrap Emergence to Composition-Sensitive Self-Evolution▼4.3k2.00
[PDF]Surprising Success, Repeated Failure: Entropy-Guided Credit Assignment for Exploration in LLM Reasoning▼4.5k2.67
[PDF]Language-Conditioned Mean-Field Unfolding: A Variational Inference Paradigm for Referring Expression Segmentation▼4.5k2.67
[PDF]X-Ragent: Tool-Augmented Reinforcement Learning for Evidence-Grounded Chest X-Ray Reasoning▼5.8k2.33
[PDF]RecAgent: Reinforcing Referring Expression Comprehension through Agentic Tool-Augmentation▼5.8k2.33
[PDF]Diminishing Returns of Data in Sparse Autoencoders: Functional Behavior does not Guarantee Feature Reproducibility▼1.2k1.67
[PDF]Capable Teacher, Myopic Scorer: Reward-Free Teacher Repair for Agentic On-Policy Distillation▲3952.67
[PDF]RelFed: Relative-Representation Federation of Heterogeneous Small Language Models for Multi-Site AIoT Sensing▼4.3k2.00
[PDF]Not Which Paramters, but Which Updates: Forgetting-Aware Fine-Tuning of Multimodal Large Language Models▲3.0k2.33
[PDF]Learn2Play Bench: How Well Do LLM Agents Learn from Experience in Unfamiliar Environments?▼4.3k2.00
[PDF]CascadeLedger: Counterfactual Auditing of Information and Resource Integrity in LLM Cascades▼1.2k1.67
[PDF]FSPO: Policy-Consistent Risk and Pareto-Feasible Control for Budgeted LLM RL Post-Training▼4.3k2.00
[PDF]WideSeek-R1: Exploring Width Scaling for Broad Information Seeking via Multi-Agent Reinforcement Learning▼4.3k2.00
[PDF]Beyond Static Interpretability: Anticipating Post-SFT Mechanisms from Pre-SFT Parameters for Better Tuning▼1.2k3.00
Withdrawn or desk rejected (1,669)
[PDF]A Window Is Not a Door: Mapping the Limits and Conditions of Knowledge Editing in Language ModelsDesk rejected2.33
[PDF]When Is a Policy Violation a Model Failure? Policy Observability and Failure Attribution in Agent-Safety EvaluationDesk rejected
[PDF]NeuroFORM: Neuro-Symbolic Formalization of Legal Rules with Counterexample-Guided RefinementWithdrawn2.00
[PDF]When to Act, Not What to Act: Prospective Authorization for Tool-Using Language ModelsWithdrawn1.33