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
191 results for “keyword:"interventions"” in ICLR 2027. 52 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Amplification Without Repair: Causal Limits of Activation Steering in Abstract Reasoning▼5.9k2.50
[PDF]ACTIV: Reducing Spatiotemporal Hallucinations in Video Language Models through Token Interventions▲9.2k2.67
[PDF]What Multilingual Models Share: The Functional and Neural Significance of Representational Convergence▲6.1k2.00
[PDF]Shared Weights, Selected Computations: How Looped Transformers Route What Each Loop Does▼4.3k2.00
[PDF]What the Nonnegative Lens Misses: Energy-Matched Causal Tests of the J-Space in Open Language Models▲11.00
[PDF]Written as a Record, Read as an Address: What a Forward Pass Leaves in an Operation's KV Cache▲3.0k2.33
[PDF]RACE: RELATION-LEVEL COUNTERFACTUAL EXPLANATIONS FOR HETEROGENEOUS GRAPH NEURAL NETWORKS▼1361.33
[PDF]Preserving What Matters: Human-Relevant Manifold Geometry of Speech Representations Under Acoustic Intervention▼5.8k2.33
[PDF]Why Do Language Models Take Shortcuts? A Mechanistic Account of Heuristic–Constraint Competition
[PDF]EviPair: Paired Behavioral Diagnostics for Evidence-Responsive Decisions in Language Models▼5.8k2.33
[PDF]CALIPER: Compute-Aware Local Intervention and Targeted Policy Updates for Tool-Augmented Agents▼1.2k1.67
[PDF]Information Anchors: Understanding and Using Forecast-Relevant Information in Time-Series Foundation Models▼5.8k2.33
[PDF]Split, Not Sparse: Ground-Truth Feature Superposition Across Scientific Domains and Model Scale▼1361.33
[PDF]Present Action Is Not Persistent Write: Causal Rights of Optimizer History in Replay-Free Continual Learning▼71.25
[PDF]The Model Already Has a Dictionary: Cumulative SVD Features as a Training-Free Alternative to Sparse Autoencoders▲3972.67
Withdrawn or desk rejected (52)
[PDF]A Window Is Not a Door: Mapping the Limits and Conditions of Knowledge Editing in Language ModelsDesk rejected2.33
[PDF]A benchmark score is not a model property: failure composition changes across operating regimesDesk rejected
[PDF]Believing Is Seeing: How VLMs Assemble Trust in an Image, and Where It Can Be ReadDesk rejected