{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4LAIMCDXBGKMSQR7IU2BCHDXAU","short_pith_number":"pith:4LAIMCDX","schema_version":"1.0","canonical_sha256":"e2c08608770994c9423f4534111c77052a2e9e833c1a94fd5d7cb6bb0daeb686","source":{"kind":"arxiv","id":"2606.21078","version":1},"attestation_state":"computed","paper":{"title":"A Validation-Gated Mechanistic Account of Suicidality Detection in LLMs","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dingjing Shi, Mike Banad, Nafiz Ahmed, Sarah Sharif","submitted_at":"2026-06-19T03:59:24Z","abstract_excerpt":"Large language models are increasingly proposed for mental-health applications such as detecting suicidal content, raising the question of what they rely on. We study this mechanistically and use it to ask a narrower question: how to make a causal claim about a model's internal features more trustworthy. Our validation-gated framework, with suicidality detection as a case study, interprets a behavior only after the model is shown to perform it: a concept is admitted only once the model ranks it above a simple lexical baseline, and each subsequent property is tested against a matched control. T"},"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":"2606.21078","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-19T03:59:24Z","cross_cats_sorted":[],"title_canon_sha256":"6a0fc84698746ed7eb8889e72f6c277eb67764504a50d8643b46a0a66d5bf780","abstract_canon_sha256":"04181460babb855a87d566ceedca60b00f1553ee9c4756efd18a40a481be2c29"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T01:12:29.112164Z","signature_b64":"glXu0Vs1Mf571wLPbpyosNQYYY+XnfCLyLuEI4JjV8K61w/Irmm7R/98mnk1PuG17nZdKfhGrUIdWMN/ZdCECQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e2c08608770994c9423f4534111c77052a2e9e833c1a94fd5d7cb6bb0daeb686","last_reissued_at":"2026-06-23T01:12:29.111723Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T01:12:29.111723Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Validation-Gated Mechanistic Account of Suicidality Detection in LLMs","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dingjing Shi, Mike Banad, Nafiz Ahmed, Sarah Sharif","submitted_at":"2026-06-19T03:59:24Z","abstract_excerpt":"Large language models are increasingly proposed for mental-health applications such as detecting suicidal content, raising the question of what they rely on. We study this mechanistically and use it to ask a narrower question: how to make a causal claim about a model's internal features more trustworthy. Our validation-gated framework, with suicidality detection as a case study, interprets a behavior only after the model is shown to perform it: a concept is admitted only once the model ranks it above a simple lexical baseline, and each subsequent property is tested against a matched control. T"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.21078","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/2606.21078/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":"2606.21078","created_at":"2026-06-23T01:12:29.111790+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.21078v1","created_at":"2026-06-23T01:12:29.111790+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.21078","created_at":"2026-06-23T01:12:29.111790+00:00"},{"alias_kind":"pith_short_12","alias_value":"4LAIMCDXBGKM","created_at":"2026-06-23T01:12:29.111790+00:00"},{"alias_kind":"pith_short_16","alias_value":"4LAIMCDXBGKMSQR7","created_at":"2026-06-23T01:12:29.111790+00:00"},{"alias_kind":"pith_short_8","alias_value":"4LAIMCDX","created_at":"2026-06-23T01:12:29.111790+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/4LAIMCDXBGKMSQR7IU2BCHDXAU","json":"https://pith.science/pith/4LAIMCDXBGKMSQR7IU2BCHDXAU.json","graph_json":"https://pith.science/api/pith-number/4LAIMCDXBGKMSQR7IU2BCHDXAU/graph.json","events_json":"https://pith.science/api/pith-number/4LAIMCDXBGKMSQR7IU2BCHDXAU/events.json","paper":"https://pith.science/paper/4LAIMCDX"},"agent_actions":{"view_html":"https://pith.science/pith/4LAIMCDXBGKMSQR7IU2BCHDXAU","download_json":"https://pith.science/pith/4LAIMCDXBGKMSQR7IU2BCHDXAU.json","view_paper":"https://pith.science/paper/4LAIMCDX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.21078&json=true","fetch_graph":"https://pith.science/api/pith-number/4LAIMCDXBGKMSQR7IU2BCHDXAU/graph.json","fetch_events":"https://pith.science/api/pith-number/4LAIMCDXBGKMSQR7IU2BCHDXAU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4LAIMCDXBGKMSQR7IU2BCHDXAU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4LAIMCDXBGKMSQR7IU2BCHDXAU/action/storage_attestation","attest_author":"https://pith.science/pith/4LAIMCDXBGKMSQR7IU2BCHDXAU/action/author_attestation","sign_citation":"https://pith.science/pith/4LAIMCDXBGKMSQR7IU2BCHDXAU/action/citation_signature","submit_replication":"https://pith.science/pith/4LAIMCDXBGKMSQR7IU2BCHDXAU/action/replication_record"}},"created_at":"2026-06-23T01:12:29.111790+00:00","updated_at":"2026-06-23T01:12:29.111790+00:00"}