{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7FVRAIFFIY74X22RZURUEIFP63","short_pith_number":"pith:7FVRAIFF","schema_version":"1.0","canonical_sha256":"f96b1020a5463fcbeb51cd234220aff6e178510351d613a94411dfeff8148605","source":{"kind":"arxiv","id":"2605.25891","version":1},"attestation_state":"computed","paper":{"title":"Causal Tongue-Tie: LLMs Can Encode Causal Direction, But Their Yes/No Outputs Fail to Express","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Xiao-Ping Zhang, Ziyi Ding","submitted_at":"2026-05-25T14:19:51Z","abstract_excerpt":"We find a mismatch between what large language models encode about a causal question and what they answer. On anti-commonsense CLadder items, a fixed linear probe recovers the evidence-supported answer from the model's hidden state (accuracy approximately 0.97), while the spoken Yes/No reverts to the commonsense one (accuracy approximately 0.5). We call this approximately +0.5 gap Causal Tongue-Tie: a wrong Yes/No decomposes into two separable failure modes: no internal signal versus a signal the verbal interface cannot say. The implication cuts both ways for output-only causal benchmarks: a b"},"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":"2605.25891","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-25T14:19:51Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3396811f7f6921fb94f919cf742dc89db49741f895d7c825462cddff8e34f700","abstract_canon_sha256":"2d13482d62dae3f6c4e492e6ef108eb012555a33d90cab02c5f4a7d1017b1bf1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:05:17.133152Z","signature_b64":"+6cXoNzK+UCHr1q4y+Xd6zYW041F5y5JQL7h8ecIvqLZMMEAJuZyleA2z97dp1Ba2GQnre35CuUBPB2pfJg/Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f96b1020a5463fcbeb51cd234220aff6e178510351d613a94411dfeff8148605","last_reissued_at":"2026-05-26T02:05:17.132660Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:05:17.132660Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Causal Tongue-Tie: LLMs Can Encode Causal Direction, But Their Yes/No Outputs Fail to Express","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Xiao-Ping Zhang, Ziyi Ding","submitted_at":"2026-05-25T14:19:51Z","abstract_excerpt":"We find a mismatch between what large language models encode about a causal question and what they answer. On anti-commonsense CLadder items, a fixed linear probe recovers the evidence-supported answer from the model's hidden state (accuracy approximately 0.97), while the spoken Yes/No reverts to the commonsense one (accuracy approximately 0.5). We call this approximately +0.5 gap Causal Tongue-Tie: a wrong Yes/No decomposes into two separable failure modes: no internal signal versus a signal the verbal interface cannot say. The implication cuts both ways for output-only causal benchmarks: a b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25891","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/2605.25891/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":"2605.25891","created_at":"2026-05-26T02:05:17.132732+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.25891v1","created_at":"2026-05-26T02:05:17.132732+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25891","created_at":"2026-05-26T02:05:17.132732+00:00"},{"alias_kind":"pith_short_12","alias_value":"7FVRAIFFIY74","created_at":"2026-05-26T02:05:17.132732+00:00"},{"alias_kind":"pith_short_16","alias_value":"7FVRAIFFIY74X22R","created_at":"2026-05-26T02:05:17.132732+00:00"},{"alias_kind":"pith_short_8","alias_value":"7FVRAIFF","created_at":"2026-05-26T02:05:17.132732+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/7FVRAIFFIY74X22RZURUEIFP63","json":"https://pith.science/pith/7FVRAIFFIY74X22RZURUEIFP63.json","graph_json":"https://pith.science/api/pith-number/7FVRAIFFIY74X22RZURUEIFP63/graph.json","events_json":"https://pith.science/api/pith-number/7FVRAIFFIY74X22RZURUEIFP63/events.json","paper":"https://pith.science/paper/7FVRAIFF"},"agent_actions":{"view_html":"https://pith.science/pith/7FVRAIFFIY74X22RZURUEIFP63","download_json":"https://pith.science/pith/7FVRAIFFIY74X22RZURUEIFP63.json","view_paper":"https://pith.science/paper/7FVRAIFF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.25891&json=true","fetch_graph":"https://pith.science/api/pith-number/7FVRAIFFIY74X22RZURUEIFP63/graph.json","fetch_events":"https://pith.science/api/pith-number/7FVRAIFFIY74X22RZURUEIFP63/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7FVRAIFFIY74X22RZURUEIFP63/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7FVRAIFFIY74X22RZURUEIFP63/action/storage_attestation","attest_author":"https://pith.science/pith/7FVRAIFFIY74X22RZURUEIFP63/action/author_attestation","sign_citation":"https://pith.science/pith/7FVRAIFFIY74X22RZURUEIFP63/action/citation_signature","submit_replication":"https://pith.science/pith/7FVRAIFFIY74X22RZURUEIFP63/action/replication_record"}},"created_at":"2026-05-26T02:05:17.132732+00:00","updated_at":"2026-05-26T02:05:17.132732+00:00"}