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
126 results for “keyword:"reproducibility"” in ICLR 2027. 36 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]The Selection Rule Decides the Winner: A Pre-Registered Audit of Open-Set Graph Anomaly Detection▼4.3k2.00
[PDF]Unstable Features, Reproducible Subspaces: Understanding Seed Dependence in Sparse Autoencoders▼5.8k2.33
[PDF]Exact, Batch-Invariant and Fast Test-Time Query Optimization for Late-Interaction Retrieval▲5.4k3.00
[PDF]Iterative distillation of Matryoshka sparse autoencoders improves hierarchical feature organization in ESMFold2▲11.00
[PDF]RAG Control Is Not Answer Fidelity: An Auditable Protocol for the Sufficiency–Fidelity Gap in Multi-Hop RAG▼1.2k1.67
[PDF]ALAMEINBENCH: EVALUATING LONG-HORIZON STRATEGIC PLANNING OF LLM AGENTS IN A HISTORICAL WARGAME▲13k2.33
[PDF]When Do Behavior Traces Validate MoE Interventions? Auditing Labels, Selection, and Decision Impact▼5.8k2.33
[PDF]Do Language and Visual Perturbations Compound? Interaction Effects and Backend Sensitivity in Compact Vision-Language-Action Model Evaluation▼1361.33
[PDF]TAIWO: Structured Test-Time Inference for Open-Ended Semantic Discovery Across Large Language Models▲11k2.00
[PDF]The Evaluator Is a Random Variable: Auditing the Fixed Feature Space Behind Text-to-Motion Benchmarks▼4.3k2.00
[PDF]Stable Accuracy, Unstable Allocations:Retraining Reproducibility in Selective Prediction▼1.2k1.67
[PDF]Scores Alone Do Not Prove Discovery: The Discovery Certification Protocol for Auditing AI Research Agents▼5.9k2.50
[PDF]When an Evaluation Rule Writes Training Labels: Measuring Human-Reference Forgiveness in NAVSIM▼5.3k1.33
[PDF]Scoring the Same Outputs: How Answer-Extractor Choice and Consistency Change Benchmark Comparisons▼1361.33
[PDF]How Much Can Sampling-Aware Uncertainty Buy? A Measurable Ceiling for Irregularly Sampled Time Series▲11.00
[PDF]Plottability Is Not Identifiability: Auditing the Retraining Stability of KAN Explanations▼1361.33
[PDF]Measure First: The Noise Floor of an Autonomous Research Loop Dominates Its Accept Rule▼5.8k2.33
[PDF]Half the Rollouts, No Measurable Cost: A Seeded Audit of Dynamic Sampling and Credit Assignment in DAPO▼1.2k1.67
[PDF]When Do Models Admit They Are Wrong? Failure Disclosure Is Unstable Under Reinforcement Learning▼1.2k1.67
[PDF]Towards Autonomous Environmental Health: AERIS, a Research Agent for Air-Pollution Exposure Assessment▼1361.33
[PDF]What Do Hypergraph Benchmarks Measure? Feature Paths, Evaluation Protocols, and Modular Constraints▼1361.33
[PDF]Which Channels, or How Many? Matched Controls for Data-Derived Attention Masks in Multivariate Forecasting▼4.3k2.00
[PDF]PGV: A Two-Check Verification Harness for Self-Play Promotion Gates, Validated on Four Real Incidents in Deployed Systems▲11.00
[PDF]How Many Samples Should a Planner Draw? The Answer Depends on a Convention Nobody Reports▲3.0k2.33
[PDF]Searching Harder Hurts at a Fixed Budget: The Unreported Candidate-Set Size in Adaptive Collocation▼1.2k1.67
[PDF]A RIGOR-MATCHED AUDIT OF PERIODIC-STEP LAYER SKIPPING FOR EFFICIENT LLM INFERENCE: CONFLAYERS VERSUS SWIFT, WITH A SUPPLEMENTAL ANALYSIS OF TRAINED ROUTING ALTERNATIVES▲11.00
[PDF]Ephemeral Models, Durable Claims: Closed-Model Dependence and the Preservation of AI Research▼5.8k2.33
[PDF]Reading the Board, Playing Another One: Turn-Level Evaluation of Open LLM Agents on Textual Sokoban▲11.00
Withdrawn or desk rejected (36)
[PDF]Beyond Execution: Auditing Experimental Fidelity in LLM-Driven Scientific ResearchDesk rejected1.33
[PDF]When Spatial Blocking Makes Distance Buffering Redundant: Structural Inertness versus Small EffectsWithdrawn1.33
[PDF]Error Bars Change the Leaderboard: Handcrafted Features Are Not Beaten in Automated Alpha MiningDesk rejected
[PDF]Candidate Likelihood Is Not Generated Stance: Auditing the Political Answers of Locally Run Open-Weight ModelsWithdrawn1.33