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
8,681 results for “keyword:"language"” in ICLR 2027. 2,080 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Beyond Flat Attribution: Recovering Boolean Evidence Formulas for Black-Box Language Models▼4.3k2.00
[PDF]Training-Free Adaptation of Generalist Robot Policies via Distributional Policy Steering▼1.3k1.75
[PDF]How Does Reasoning Become Implicit? Uncovering the Dynamics of Chain-of-Thought Internalization▲3912.67
[PDF]When Unlearning Is Not the End: Controlled Oracle Recovery Against Federated Unlearning▼1.2k1.67
[PDF]Learning to Reason from Reference Samples Alone: Reinforced Adversarial Learning for Modern LLM Post-Training▲15k3.33
[PDF]GenTell-Gaze: A Dataset and Benchmark for Localising Physical Anomalies in AI-Generated Video with Human Gaze▲5103.00
[PDF]Beyond Rejection Sampling: Learning from Guided Rollouts to Improve LLMs for Software Engineering▼4.3k2.00
[PDF]Phase-MoE: A Co-Design to Bound the Expert Explosion in Block Diffusion Language Models▼4.3k2.00
[PDF]ModelParts: Model-Grounded Tasks and Roofline Rewards for Training Kernel Agents on Emerging Accelerators▼4.3k2.00
Withdrawn or desk rejected (2,080)
[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]Population Contracts for Scalable Language-Guided Multi-Robot Planning and ControlDesk rejected