{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:M7O7MGVLNYFCUKG3ZOIAY4QUCW","short_pith_number":"pith:M7O7MGVL","schema_version":"1.0","canonical_sha256":"67ddf61aab6e0a2a28dbcb900c7214158d0417fcb2bb5a563d919077ad640e4d","source":{"kind":"arxiv","id":"2607.02140","version":1},"attestation_state":"computed","paper":{"title":"Probing Chemical Language Models: Effects of Pre-training and Fine-tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anna Karnysheva, Dietrich Klakow, Ji-Ung Lee","submitted_at":"2026-07-02T13:16:31Z","abstract_excerpt":"Chemical language models (CLMs) are trained with linearized representations such as SMILES, yet it remains unclear which chemically meaningful substructures they encode. To foster a better understanding of CLMs, we conduct a systematic study and probe for 78 molecular substructures across eight pre-trained and six randomly initialized models. We furthermore study how fine-tuning on chemical downstream tasks affects the learned representations of molecular substructures. Our results show that pre-training generally improves molecular structure awareness of CLMs, particularly in the upper layers"},"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":"2607.02140","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-02T13:16:31Z","cross_cats_sorted":[],"title_canon_sha256":"bf2ff6e97e830afc9c064d9b81dea761dfdae294d82e16e524cc85c417401db1","abstract_canon_sha256":"2e4b0223a4921f74124923ee6d84f3980ea1ae2144db610d90a4695cb20ccf46"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T01:17:43.233766Z","signature_b64":"wcYN5y4f76if0aY4L6QYYR83+gByCF5QlxDYpuSLMVMPFS6Fj3zLKKfq0CZfgSl8y37QoFcmVRpg5gbCueAPDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"67ddf61aab6e0a2a28dbcb900c7214158d0417fcb2bb5a563d919077ad640e4d","last_reissued_at":"2026-07-03T01:17:43.233335Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T01:17:43.233335Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Probing Chemical Language Models: Effects of Pre-training and Fine-tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anna Karnysheva, Dietrich Klakow, Ji-Ung Lee","submitted_at":"2026-07-02T13:16:31Z","abstract_excerpt":"Chemical language models (CLMs) are trained with linearized representations such as SMILES, yet it remains unclear which chemically meaningful substructures they encode. To foster a better understanding of CLMs, we conduct a systematic study and probe for 78 molecular substructures across eight pre-trained and six randomly initialized models. We furthermore study how fine-tuning on chemical downstream tasks affects the learned representations of molecular substructures. Our results show that pre-training generally improves molecular structure awareness of CLMs, particularly in the upper layers"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.02140","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/2607.02140/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":"2607.02140","created_at":"2026-07-03T01:17:43.233400+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.02140v1","created_at":"2026-07-03T01:17:43.233400+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.02140","created_at":"2026-07-03T01:17:43.233400+00:00"},{"alias_kind":"pith_short_12","alias_value":"M7O7MGVLNYFC","created_at":"2026-07-03T01:17:43.233400+00:00"},{"alias_kind":"pith_short_16","alias_value":"M7O7MGVLNYFCUKG3","created_at":"2026-07-03T01:17:43.233400+00:00"},{"alias_kind":"pith_short_8","alias_value":"M7O7MGVL","created_at":"2026-07-03T01:17:43.233400+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/M7O7MGVLNYFCUKG3ZOIAY4QUCW","json":"https://pith.science/pith/M7O7MGVLNYFCUKG3ZOIAY4QUCW.json","graph_json":"https://pith.science/api/pith-number/M7O7MGVLNYFCUKG3ZOIAY4QUCW/graph.json","events_json":"https://pith.science/api/pith-number/M7O7MGVLNYFCUKG3ZOIAY4QUCW/events.json","paper":"https://pith.science/paper/M7O7MGVL"},"agent_actions":{"view_html":"https://pith.science/pith/M7O7MGVLNYFCUKG3ZOIAY4QUCW","download_json":"https://pith.science/pith/M7O7MGVLNYFCUKG3ZOIAY4QUCW.json","view_paper":"https://pith.science/paper/M7O7MGVL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.02140&json=true","fetch_graph":"https://pith.science/api/pith-number/M7O7MGVLNYFCUKG3ZOIAY4QUCW/graph.json","fetch_events":"https://pith.science/api/pith-number/M7O7MGVLNYFCUKG3ZOIAY4QUCW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M7O7MGVLNYFCUKG3ZOIAY4QUCW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M7O7MGVLNYFCUKG3ZOIAY4QUCW/action/storage_attestation","attest_author":"https://pith.science/pith/M7O7MGVLNYFCUKG3ZOIAY4QUCW/action/author_attestation","sign_citation":"https://pith.science/pith/M7O7MGVLNYFCUKG3ZOIAY4QUCW/action/citation_signature","submit_replication":"https://pith.science/pith/M7O7MGVLNYFCUKG3ZOIAY4QUCW/action/replication_record"}},"created_at":"2026-07-03T01:17:43.233400+00:00","updated_at":"2026-07-03T01:17:43.233400+00:00"}