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
375 results for “keyword:"steering"” in ICLR 2027. 78 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Anchored but Not Protected: When Refusal Erodes During Fine-Tuning, and Whether Holding One Direction Stops It▲13k2.33
[PDF]INTERVENTION IS NOT IDENTIFICATION: ACTIVATION EDITS CAN PRODUCE THE PREDICTED EFFECT FOR THE WRONG REASON▼1361.33
[PDF]CircuitSteer: Geometrically Aligned Multi-Layer Steering via Sparse Autoencoder Circuits▼4.5k2.67
[PDF]MAPP: Cross-Lingual Transfer through Representation Alignment with Performance-Gap Probing▲3.0k2.33
[PDF]One Token Can Be Enough: Bridging Prompting and Activation Steering with Prefix Steering▲5103.00
[PDF]Constitutional Value Potentials: Reading Priority Margins and Steering Value Trade-offs in Language-Model Activations▲3.0k2.33
[PDF]Hesitation Has a Geometry: Entropy-Trained Hyperbolic Probes for Sparse Activation Steering▼4.3k2.00
[PDF]Training-Free Adaptation of Generalist Robot Policies via Distributional Policy Steering▼1.3k1.75
Withdrawn or desk rejected (78)
[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
[PDF]Meaning Beyond Language: Learning Complementary Semantics for Multimodal Metaphor DetectionWithdrawn1.33