{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TNBDJFV7HKDGLFASOSRKSVU33K","short_pith_number":"pith:TNBDJFV7","schema_version":"1.0","canonical_sha256":"9b423496bf3a8665941274a2a9569bda84990584f32574996ee4c721cd7909b8","source":{"kind":"arxiv","id":"2606.04915","version":1},"attestation_state":"computed","paper":{"title":"Caliper: Probing Lexical Anchors versus Causal Structure in LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IR"],"primary_cat":"cs.CL","authors_text":"Shuigeng Zhou, Zhenyu Yu","submitted_at":"2026-06-03T14:11:16Z","abstract_excerpt":"Large language models reach 50 to 70% accuracy on causal reasoning benchmarks such as CLadder, but it is unclear whether this reflects structural reasoning or lexical pattern matching. We introduce Caliper, a controlled perturbation that replaces semantic variable names with placeholder tokens while preserving the causal graph and probabilistic specification of each question. Across nine instruction-tuned LLMs from 3.8B to 671B and three causal reasoning benchmarks, lexical anonymization yields robust accuracy drops of +7.6, +27.0, and +11.1 pp on a local 3.8B-14B set, rising to +29.6 and +18."},"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":"2606.04915","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-06-03T14:11:16Z","cross_cats_sorted":["cs.IR"],"title_canon_sha256":"5029ddb83d65e259e379c9412f14eb5eb11165c5322d6390ba0787960a77e052","abstract_canon_sha256":"3b7385f737de93f772dae5cc8688719f161d88a57b5109e81e26d52512a8219b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:09:55.336647Z","signature_b64":"NPL0uwva/QTyhJ71YiBcP8Wk7eiVPPARwviLhV/v00w8lV/zl+O42BAxvexpeEP4vMXOnGc1VbKFDGYQuheOBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9b423496bf3a8665941274a2a9569bda84990584f32574996ee4c721cd7909b8","last_reissued_at":"2026-06-04T01:09:55.335940Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:09:55.335940Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Caliper: Probing Lexical Anchors versus Causal Structure in LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IR"],"primary_cat":"cs.CL","authors_text":"Shuigeng Zhou, Zhenyu Yu","submitted_at":"2026-06-03T14:11:16Z","abstract_excerpt":"Large language models reach 50 to 70% accuracy on causal reasoning benchmarks such as CLadder, but it is unclear whether this reflects structural reasoning or lexical pattern matching. We introduce Caliper, a controlled perturbation that replaces semantic variable names with placeholder tokens while preserving the causal graph and probabilistic specification of each question. Across nine instruction-tuned LLMs from 3.8B to 671B and three causal reasoning benchmarks, lexical anonymization yields robust accuracy drops of +7.6, +27.0, and +11.1 pp on a local 3.8B-14B set, rising to +29.6 and +18."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.04915","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/2606.04915/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":"2606.04915","created_at":"2026-06-04T01:09:55.336044+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.04915v1","created_at":"2026-06-04T01:09:55.336044+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.04915","created_at":"2026-06-04T01:09:55.336044+00:00"},{"alias_kind":"pith_short_12","alias_value":"TNBDJFV7HKDG","created_at":"2026-06-04T01:09:55.336044+00:00"},{"alias_kind":"pith_short_16","alias_value":"TNBDJFV7HKDGLFAS","created_at":"2026-06-04T01:09:55.336044+00:00"},{"alias_kind":"pith_short_8","alias_value":"TNBDJFV7","created_at":"2026-06-04T01:09:55.336044+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/TNBDJFV7HKDGLFASOSRKSVU33K","json":"https://pith.science/pith/TNBDJFV7HKDGLFASOSRKSVU33K.json","graph_json":"https://pith.science/api/pith-number/TNBDJFV7HKDGLFASOSRKSVU33K/graph.json","events_json":"https://pith.science/api/pith-number/TNBDJFV7HKDGLFASOSRKSVU33K/events.json","paper":"https://pith.science/paper/TNBDJFV7"},"agent_actions":{"view_html":"https://pith.science/pith/TNBDJFV7HKDGLFASOSRKSVU33K","download_json":"https://pith.science/pith/TNBDJFV7HKDGLFASOSRKSVU33K.json","view_paper":"https://pith.science/paper/TNBDJFV7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.04915&json=true","fetch_graph":"https://pith.science/api/pith-number/TNBDJFV7HKDGLFASOSRKSVU33K/graph.json","fetch_events":"https://pith.science/api/pith-number/TNBDJFV7HKDGLFASOSRKSVU33K/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TNBDJFV7HKDGLFASOSRKSVU33K/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TNBDJFV7HKDGLFASOSRKSVU33K/action/storage_attestation","attest_author":"https://pith.science/pith/TNBDJFV7HKDGLFASOSRKSVU33K/action/author_attestation","sign_citation":"https://pith.science/pith/TNBDJFV7HKDGLFASOSRKSVU33K/action/citation_signature","submit_replication":"https://pith.science/pith/TNBDJFV7HKDGLFASOSRKSVU33K/action/replication_record"}},"created_at":"2026-06-04T01:09:55.336044+00:00","updated_at":"2026-06-04T01:09:55.336044+00:00"}