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
689 results for “keyword:"mechanistic interpretability"” in ICLR 2027. 142 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]SharedSAE: A Unified Sparse Dictionary across U-Net Blocks and Denoising Time for Diffusion Models▼4.3k2.00
[PDF]Overtopping in LLMs: Spiking-Like Causal Control and the Limits of Static Intervention▼5.8k2.33
[PDF]Circuit Emergence Does Not Establish Functional Emergence: Reasoning Distillation Can Recruit Latent Attention-Head Signals▼5.8k2.33
[PDF]Active Budget Can Kill Sensitivity: Diagnosing and Repairing TopK Sparse Autoencoder Reliability▼4.3k2.00
[PDF]What Do Activation Interventions Establish? Causal Controls for Autoregressive Time-Series Foundation Models▼4.3k2.00
[PDF]Content-Dependent Recovery from Task-Relevant Activation Disruption in Large Language Models▲3952.67
[PDF]The Attribution Blind Spot: Layerwise Trajectory Diagnostics for Source Reliance in Retrieval-Augmented Language Models▼4.3k2.00
[PDF]The Missing Semantic Reference: Consensus Geometry and Observable Dynamics in Language Models▼4.5k2.67
[PDF]Transformers track Bayesian evidence for latent common causes via a context-invariant mechanism▲3.0k2.33
[PDF]They Know Where It Points, But Not What It Means: The Causal Grammar of Relational Visual Prompts▼4.3k2.00
[PDF]Beyond Linear Concepts: Discovering and Aligning Non-Linear Concept Manifolds in Large Language Models▼4.5k2.67
[PDF]Improving the Robustness of VLAs to Spatial Object Perturbation via Attention-Head Amplification▼3.7k2.25
[PDF]When Speed Costs Accuracy: Causal Faithfulness of Component Attribution for Gender-Bias Localization▼1361.33
[PDF]Dictionary Size and Dead Latents Handling Jointly Set the Recovery-Optimal Sparsity of Top-K Sparse Autoencoders▼1361.33
[PDF]From Embeddings to Attention Heads: Input-Agnostic Tracing of Embedding-MLP-Attention Circuits in LLMs▼4.3k2.00
[PDF]Beyond Pattern Matching: Tracing Symbolic Reasoning Failures in LLMs to Their Mechanistic Origin▼5.8k2.33
[PDF]Deep latents, long memory: Latent persistence explains temporal correlations in transformer activations▲3952.67
Withdrawn or desk rejected (142)
[PDF]Why Benign Post-Training Induces Safety Gate Collapse: Changes in Harmful Activation Geometry Relative to the Inherited Refusal GateWithdrawn1.67
[PDF]Concept Subspaces Compute Beyond the Logit Lens: A Weights-Only Test for Locating Representations Upstream of ReadoutWithdrawn