{"version":1,"type":"story","url":"https://digestai.news/story/circuit-oracle-automates-attribution-graph-analysis","json":"https://digestai.news/story/circuit-oracle-automates-attribution-graph-analysis.json","markdown":"https://digestai.news/story/circuit-oracle-automates-attribution-graph-analysis.md","slug":"circuit-oracle-automates-attribution-graph-analysis","headline":"Circuit Oracle automates attribution graph analysis","summary":"Circuit Oracle is an off-the-shelf LLM designed to automate the interpretation of language model computations. It does this by analyzing attribution graphs, which are used for mechanistic interpretability. The tool can answer questions about a target model through tool calls and has been tested on three safety-relevant tasks: detecting spurious features in probe circuits, eliciting hidden knowledge from taboo-finetuned models, and suppression jailbreaking via feature-level interventions. Results show mixed performance; Circuit Oracle outperforms single-layer feature-ranking baselines in spurious-feature detection but remains below self-explainers like Activation Oracles in secret elicitation tasks. The tool produces pruned subgraphs and context-specific feature explanations that are qualitatively assessed.","keyPoints":["Circuit Oracle automates attribution graph analysis","Tests on three safety-relevant proxy tasks mixed","Outperforms single-layer feature-ranking baselines in spurious-feature detection"],"whyItMatters":"Automating the interpretation of language model computations can lead to more efficient and accurate insights, potentially improving interpretability and security.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["Circuit Oracle"],"models":[],"people":[]},"firstPublishedAt":"2026-11-03T05:00:00Z","updatedAt":"2026-11-03T05:00:00Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"arXiv cs.CL","title":"Circuit Oracle: Automating Attribution Graph Analysis via Natural-Language Queries","url":"https://arxiv.org/abs/2611.00420","publishedAt":"2026-11-03T05:00:00Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Circuit Oracle automates attribution graph analysis\", 3 November 2026, https://digestai.news/story/circuit-oracle-automates-attribution-graph-analysis","publisher":"Digest AI","title":"Circuit Oracle automates attribution graph analysis","datePublished":"2026-11-03T05:00:00Z","url":"https://digestai.news/story/circuit-oracle-automates-attribution-graph-analysis"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}