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
75 results for “keyword:"off policy"” in ICLR 2027. 9 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Reweight, Do Not Re-calibrate: Reusing the Same Encoder-Predictor Pair (World Model) When the Policy or Query Changes▼1361.33
[PDF]Preserving Every Target Row Does Not Preserve Propagated Uncertainty in Recursive Fitted Evaluation▼4.5k2.67
[PDF]What Off-Policy Corrections Keep: Gradient Direction and Realized Gains from Stale LLM Rollouts▲5.4k3.00
[PDF]Fine-Tuning on Self-Generated and Reward-Weighted Data: Learning Dynamcis, Convergence Rates, and Benefits of Off-Policyness▲153.33
[PDF]SHARV: A Simple Multi-Horizon Approach to Representation and Value Learning in Reinforcement Learning▼4.5k2.67
[PDF]Offline Policy Evaluation via Mixed Bellman Residuals and Adaptive Critic Representations▲14k2.50
[PDF]Learning without Overwriting: A Theory of Self-Distillation and Supervised Fine-Tuning in Continual Reasoning▲6.8k3.67
Withdrawn or desk rejected (9)
[PDF]When Agents Change Their Own Test: Prospective Target Preservation under Candidate-Induced Task ExpansionWithdrawn1.33
[PDF]AsyncWebRL: Efficient Asynchronous Reinforcement Learning for Multi-Step Visual Web AgentsWithdrawn