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
22 results for “keyword:"evaluation benchmarks"” in ICLR 2027. 15 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]StateTrackBench: Executable Trajectories for Locating Where Language Models Lose Track of Evolving State▲11.00
[PDF]RHContrast: A Matched-Control Benchmark for Evaluating Reward-Hacking Risks in Long-Form LLM Responses▲3.0k2.33
[PDF]Does Better Reprojection Mean Better 3D Reconstruction? Candidate- and Evidence-Stratified Evaluation for FLFM▼1.2k1.67
[PDF]CoHabit Arena: Evaluating Personal Preference Realization in Long-Horizon Human-Agent CoHabitation▼1.2k1.67
Withdrawn or desk rejected (15)
[PDF]Reference-Dependent Scores, Margin-Stable Decisions: Calibrating MRI Reconstruction BenchmarksWithdrawn1.67
[PDF]Beyond Isolated Clips: Benchmarking Inter-Segment Narrative Continuity in Movie-Level Video UnderstandingWithdrawn1.33
[PDF]Feasible-Source Rules Define Dataset-Mixture Evaluation Identity: Cross-Schema Fixed-Pool InterventionsDesk rejected
[PDF]Matched Parent-Removal Audits Expose When Augmentation Provenance Changes DecisionsDesk rejected