SOURCE-LINKED INTELLIGENCE

Circuit Condensation: Post-Training that Concentrates a Behavior's Causal Circuit

arXiv · AI, language, vision and robotics · article · Aug 27, 2026 · UTC

One approach to mechanistic interpretability explains behavior through circuits: the components and connections that carry it. Frozen discovery often returns hundreds of edges, making them hard to inspect, compare, or verify exhaustively. We introduce Circuit Condensation, which post-trains models to concentrate behaviors into smaller causal graphs. Each round prunes low-attribution edges and trains a low-rank adapter to match the original through what remains, retaining the cut only if task performance and general capability survive. Across four behaviors and eight models, condensed circuits

Read original source ↗ Open in workspace

recordType

paper

region

Global