{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:SUXZN6UGQQFPVEY5VWIBFRFFUU","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"39c096003002cd5e9532b00d02145c74ae2b5de2e77f2a62e8e690a67464edad","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T17:25:37Z","title_canon_sha256":"2438d7cf21176adec9d791d6429028b31e6e72efb89dd512d3d945d03b0e6573"},"schema_version":"1.0","source":{"id":"2402.17700","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.17700","created_at":"2026-07-05T08:59:26Z"},{"alias_kind":"arxiv_version","alias_value":"2402.17700v2","created_at":"2026-07-05T08:59:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.17700","created_at":"2026-07-05T08:59:26Z"},{"alias_kind":"pith_short_12","alias_value":"SUXZN6UGQQFP","created_at":"2026-07-05T08:59:26Z"},{"alias_kind":"pith_short_16","alias_value":"SUXZN6UGQQFPVEY5","created_at":"2026-07-05T08:59:26Z"},{"alias_kind":"pith_short_8","alias_value":"SUXZN6UG","created_at":"2026-07-05T08:59:26Z"}],"graph_snapshots":[{"event_id":"sha256:f6dcba735e28327f0f2257336c3c4f0642cacaa1057b07de68a9f4dba99adcdc","target":"graph","created_at":"2026-07-05T08:59:26Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2402.17700/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Individual neurons participate in the representation of multiple high-level concepts. To what extent can different interpretability methods successfully disentangle these roles? To help address this question, we introduce RAVEL (Resolving Attribute-Value Entanglements in Language Models), a dataset that enables tightly controlled, quantitative comparisons between a variety of existing interpretability methods. We use the resulting conceptual framework to define the new method of Multi-task Distributed Alignment Search (MDAS), which allows us to find distributed representations satisfying multi","authors_text":"Atticus Geiger, Christopher Potts, Jing Huang, Mor Geva, Zhengxuan Wu","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T17:25:37Z","title":"RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.17700","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:0880e0c1fd2c52e706a584ff48fc3900e48639cd145791b31983fa1ca981b90c","target":"record","created_at":"2026-07-05T08:59:26Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"39c096003002cd5e9532b00d02145c74ae2b5de2e77f2a62e8e690a67464edad","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T17:25:37Z","title_canon_sha256":"2438d7cf21176adec9d791d6429028b31e6e72efb89dd512d3d945d03b0e6573"},"schema_version":"1.0","source":{"id":"2402.17700","kind":"arxiv","version":2}},"canonical_sha256":"952f96fa86840afa931dad9012c4a5a51470c7a3c936a257ec7d9d642736e6f9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"952f96fa86840afa931dad9012c4a5a51470c7a3c936a257ec7d9d642736e6f9","first_computed_at":"2026-07-05T08:59:26.357237Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:59:26.357237Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YqQfo3anrPIL7GPpwoUc8ECBFZ6Th8juT3TvVdjamP0lAFvJUV82ADcCG/sj0fV1X3/ZnOOzlOQR52Q/0UpSCw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:59:26.357732Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.17700","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0880e0c1fd2c52e706a584ff48fc3900e48639cd145791b31983fa1ca981b90c","sha256:f6dcba735e28327f0f2257336c3c4f0642cacaa1057b07de68a9f4dba99adcdc"],"state_sha256":"47b2dda77a9981bf79cc14201251bd98b433d4785a97918157d776314693c154"}