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
1,261 results for “keyword:"interpretability"” in ICLR 2027. 264 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Sparse AutoEncoders Reveal That MoE Experts Exhibit Human-Interpretable Fine-Grained Specialization▼1.2k1.67
[PDF]Split, Not Sparse: Ground-Truth Feature Superposition Across Scientific Domains and Model Scale▼1361.33
[PDF]Where You Steer Shapes the Direction: Operator Conditioned LLM Behavior Control and J-lens Interpretation▲3952.67
[PDF]SPECTRAL SHAPE RESIDUAL: A LIGHTWEIGHT SPECTRAL DIAGNOSTIC FOR ATTENTION WEIGHT MATRICES▼1361.33
[PDF]Why This Expert? A Feature-Level Analysis of Routing in Mixture-of-Experts with Sparse Autoencoders▼1.2k1.67
[PDF]RefusalTrace: Localizing and Validating the Internal Refusal Pathway of LLMs with Sparse Autoencoder Features▲9.7k2.75
[PDF]A Memory That Only Looks Like One: A Hybrid Vision-Language Model Relays the Picture Rather Than Storing It▼1.2k1.67
[PDF]The Geometry of Contextual Relations: Language Models Address Facts by Order of Mention▼4.3k2.00
[PDF]Targeted Retrieval, Compact Representations: How CoT Reasoning Improves Long-Context Counting▼4.3k2.00
[PDF]Evaluating Semantic State Response and Generated Visual Feedback under Distinct Input Contracts▼1361.33
[PDF]When Weak Features Hide: Separating Statistical Detection from Sparse Autoencoder Recovery▲6.1k2.00
[PDF]Recoverable but Unreadable: Exact Accessibility Laws in Noisy Latent Parity Codes and Beyond▲5103.00
[PDF]Do Language Models Have a Model of You? Discovering and Enhancing Latent Representations of User Personas▼4.3k2.00
Withdrawn or desk rejected (264)
[PDF]A Window Is Not a Door: Mapping the Limits and Conditions of Knowledge Editing in Language ModelsDesk rejected2.33