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
11,747 results for “keyword:"models"” in ICLR 2027. 2,924 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Empathic Accommodation: LLMs Skew Information to Favor Users’ Inferred Emotional Stakes from Role Cues▼1.2k1.67
[PDF]Freeze the Decoder, Heal the Encoder: Parameter-Efficient Adaptation for SVD-Based KV-Cache Compression▼1351.33
[PDF]What Makes a Generated Cell Realistic? Dissecting Biological Fidelity in Single-Cell Generative Models▲3762.67
[PDF]EViC-VL: Evidence-Conditioned Verification for Confident Object Hallucinations in Vision-Language Models▼1351.33
[PDF]What Survives Extreme Compression? Vision-Language Embedding Alignment at Microcontroller Scale▼1.2k1.67
[PDF]Hearing Anywhere from Anything: Geometry-Conditioned Acoustic Field Synthesis from Unlocalized Blind Recordings▲3762.67
[PDF]Confidence Is Not Coverage: Budgeted Confidence Ranking Can Fail Under Fixed Reconstruction Budgets▼3.7k2.25
[PDF]DynRule: Benchmarking and Modeling Dynamic Rule Interactions for Universal Multimodal Content Moderation▼5.9k2.33
[PDF]Beyond a Better Score: Long-Horizon Agentic ML Development and Evaluation Protocol for Physics Time Series▲5.4k3.00
[PDF]TOFU: Fixing Supervised Fine-Tuning induced diversity collapse in Large Language Models▲3762.67
Withdrawn or desk rejected (2,924)
[PDF]Mind Your Marginals: What Governs Fidelity in Linearized Hierarchical Optimal Transport?Desk rejected1.00
[PDF]A Window Is Not a Door: Mapping the Limits and Conditions of Knowledge Editing in Language ModelsDesk rejected2.33
[PDF]Propagation Is Not Decoration: When Physical Diffraction Carries Generative ComputationWithdrawn1.00
[PDF]When Is a Policy Violation a Model Failure? Policy Observability and Failure Attribution in Agent-Safety EvaluationDesk rejected