{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UWTXBUR3XGXKA4SU2GLUFY73O2","short_pith_number":"pith:UWTXBUR3","schema_version":"1.0","canonical_sha256":"a5a770d23bb9aea07254d19742e3fb7685ead2f74bde80be9e5a01c5fe178fc5","source":{"kind":"arxiv","id":"2607.20379","version":1},"attestation_state":"computed","paper":{"title":"Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Hiskias Dingeto","submitted_at":"2026-07-22T17:10:23Z","abstract_excerpt":"Natural-language autoencoders score explanations of hidden activations by reconstruction: an explanation is deemed faithful if the activation can be regenerated from it. The test is structurally insensitive to individual false claims: if flipping a claim does not change the reconstruction, the claim is never penalized. We show the test is passed in two ways, neither faithful. On a released Qwen-2.5-7B verbalizer, explanations reconstruct well above chance while ~2% of specific claims are reconstruction-dependent, so the score tracks gist, not specific facts. Under exact synthetic ground truth,"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2607.20379","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-22T17:10:23Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"1d2e07041f2feb4aa0bfaa1df129289e7adbd77e2bfa1bcc5655413ccd242d28","abstract_canon_sha256":"2d74d1ce60351ba4d65b787712922c4a9e687cafbb07fd1634e1786d751c9835"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-23T01:25:17.710980Z","signature_b64":"Phj3FWlJsP7wb0tNaFzx476CtWJY4TwVWSeSN47oVYH1ycwVzuke6g165XfwxwQuHmY4NCsWNL2Pu9hvNLvxCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a5a770d23bb9aea07254d19742e3fb7685ead2f74bde80be9e5a01c5fe178fc5","last_reissued_at":"2026-07-23T01:25:17.710095Z","signature_status":"signed_v1","first_computed_at":"2026-07-23T01:25:17.710095Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Hiskias Dingeto","submitted_at":"2026-07-22T17:10:23Z","abstract_excerpt":"Natural-language autoencoders score explanations of hidden activations by reconstruction: an explanation is deemed faithful if the activation can be regenerated from it. The test is structurally insensitive to individual false claims: if flipping a claim does not change the reconstruction, the claim is never penalized. We show the test is passed in two ways, neither faithful. On a released Qwen-2.5-7B verbalizer, explanations reconstruct well above chance while ~2% of specific claims are reconstruction-dependent, so the score tracks gist, not specific facts. Under exact synthetic ground truth,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.20379","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2607.20379/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2607.20379","created_at":"2026-07-23T01:25:17.710564+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.20379v1","created_at":"2026-07-23T01:25:17.710564+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.20379","created_at":"2026-07-23T01:25:17.710564+00:00"},{"alias_kind":"pith_short_12","alias_value":"UWTXBUR3XGXK","created_at":"2026-07-23T01:25:17.710564+00:00"},{"alias_kind":"pith_short_16","alias_value":"UWTXBUR3XGXKA4SU","created_at":"2026-07-23T01:25:17.710564+00:00"},{"alias_kind":"pith_short_8","alias_value":"UWTXBUR3","created_at":"2026-07-23T01:25:17.710564+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2","json":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2.json","graph_json":"https://pith.science/api/pith-number/UWTXBUR3XGXKA4SU2GLUFY73O2/graph.json","events_json":"https://pith.science/api/pith-number/UWTXBUR3XGXKA4SU2GLUFY73O2/events.json","paper":"https://pith.science/paper/UWTXBUR3"},"agent_actions":{"view_html":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2","download_json":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2.json","view_paper":"https://pith.science/paper/UWTXBUR3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.20379&json=true","fetch_graph":"https://pith.science/api/pith-number/UWTXBUR3XGXKA4SU2GLUFY73O2/graph.json","fetch_events":"https://pith.science/api/pith-number/UWTXBUR3XGXKA4SU2GLUFY73O2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2/action/storage_attestation","attest_author":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2/action/author_attestation","sign_citation":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2/action/citation_signature","submit_replication":"https://pith.science/pith/UWTXBUR3XGXKA4SU2GLUFY73O2/action/replication_record"}},"created_at":"2026-07-23T01:25:17.710564+00:00","updated_at":"2026-07-23T01:25:17.710564+00:00"}