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
215 results for “keyword:"failure"” in ICLR 2027. 69 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Which Conditions Matter? Task-Conditioned Criticality for Task Plan Veriffcation and Repair▼4.3k2.00
[PDF]Doubt Before Acting: Assumption-Level Representation Editing for Safe Embodied LLM Planning▲5.9k1.67
[PDF]Learning What vs. Learning How: Decomposing Task and Interface Competence in Agentic Training▲5103.00
[PDF]Minimal Repair Subtrajectories: Verifier-Certified Local Recovery for Long-Horizon Agents▼1.2k1.67
[PDF]Training-Free Failure Attribution in Multi-Agent Systems via Structured Trajectory Diagnosis▼1351.33
[PDF]When Correct Plans Lead to Wrong Actions: Multi-Source Evidence Fabrication Against LLM Agents▼4.3k2.00
[PDF]Not Proven Is Not Proven-Not: Language Models Read the “Not” in a Question as Negation as Failure▼4.3k2.00
[PDF]Where Does Reasoning Fail? Attributing Distributed LLM Failures through Critical Evidence▼1.2k1.67
[PDF]Reliable Virtual Sensing: A Multi-Domain Benchmark for Robustness Under Sensor Failures▼5.8k2.33
[PDF]Out of Step with Evidence: Task Progress State Mismatch in Long-Horizon Tool-Use Agents▼1.2k1.67
[PDF]Spectral Unit Tests for Predictive Representation Models: A Fault Atlas and Its Scope Conditions▲4002.67
[PDF]biCurriculum: Crediting Executable Task and Environment Rewrites by Their Learning Utility▼1351.33
[PDF]Reliable and Efficient Agent Loops for Reinforcement Learning on Software Engineering Tasks▲6.1k2.00
[PDF]When Do Models Admit They Are Wrong? Failure Disclosure Is Unstable Under Reinforcement Learning▼1.2k1.67
[PDF]Root-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent Failures▲6.1k2.00
[PDF]Reading the Board, Playing Another One: Turn-Level Evaluation of Open LLM Agents on Textual Sokoban▲11.00
Withdrawn or desk rejected (69)
[PDF]When Is a Policy Violation a Model Failure? Policy Observability and Failure Attribution in Agent-Safety EvaluationDesk rejected
[PDF]The Repair Is Not the Task: Execution Records and False Completion in Agent RecoveryWithdrawn1.33
[PDF]When Does an Observable Lose Failure Evidence? Exact Ranking Losses for Runtime Detection in Vision–Language–Action PoliciesWithdrawn1.00
[PDF]A benchmark score is not a model property: failure composition changes across operating regimesDesk rejected