{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6C7TQIUPWFIR2VT7WWK73RKDDT","short_pith_number":"pith:6C7TQIUP","schema_version":"1.0","canonical_sha256":"f0bf38228fb1511d567fb595fdc5431cf35dff8565a8f93a9235412fe4d2b1fc","source":{"kind":"arxiv","id":"2606.16517","version":2},"attestation_state":"computed","paper":{"title":"How Post-Training Shapes Biological Reasoning Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["q-bio.QM"],"primary_cat":"cs.LG","authors_text":"Bryan Perozzi, Eric Wang, Hanlin Zhang, Lukas Fesser, Marinka Zitnik, Michelle M. Li, Sham M. Kakade, Shekoofeh Azizi","submitted_at":"2026-06-15T10:19:49Z","abstract_excerpt":"Scientific reasoning models for biology combine language models with foundation models trained on multimodal biological data, including DNA, RNA, and proteins. These models are built through post-training, yet how each stage shapes reasoning and generalization remains poorly understood. We study when post-training improves performance and when it induces over-specialization. Across genomics, transcriptomics, and proteins, we train and evaluate more than 100 biological reasoning models under controlled variation in backbone, continued pre-training (CPT), supervised fine-tuning (SFT), and reinfo"},"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":"2606.16517","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-15T10:19:49Z","cross_cats_sorted":["q-bio.QM"],"title_canon_sha256":"1a89e7dedc05edd50d117cc7df488e444c98444045b599cd53d4dfb502c80e70","abstract_canon_sha256":"095dc2cbce0f613b432b59368b2476971d040596f449b1c1c709a1461aad0e5f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-01T01:17:53.017943Z","signature_b64":"47508fcNJlamvS+GwwTjVFRokDPU/VbW+qhjlgne26V896cefxAev+E+ODoKQhFRXAqHtP/Jk3wyRco9eO9NBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f0bf38228fb1511d567fb595fdc5431cf35dff8565a8f93a9235412fe4d2b1fc","last_reissued_at":"2026-07-01T01:17:53.017533Z","signature_status":"signed_v1","first_computed_at":"2026-07-01T01:17:53.017533Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Post-Training Shapes Biological Reasoning Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["q-bio.QM"],"primary_cat":"cs.LG","authors_text":"Bryan Perozzi, Eric Wang, Hanlin Zhang, Lukas Fesser, Marinka Zitnik, Michelle M. Li, Sham M. Kakade, Shekoofeh Azizi","submitted_at":"2026-06-15T10:19:49Z","abstract_excerpt":"Scientific reasoning models for biology combine language models with foundation models trained on multimodal biological data, including DNA, RNA, and proteins. These models are built through post-training, yet how each stage shapes reasoning and generalization remains poorly understood. We study when post-training improves performance and when it induces over-specialization. Across genomics, transcriptomics, and proteins, we train and evaluate more than 100 biological reasoning models under controlled variation in backbone, continued pre-training (CPT), supervised fine-tuning (SFT), and reinfo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.16517","kind":"arxiv","version":2},"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/2606.16517/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":"2606.16517","created_at":"2026-07-01T01:17:53.017587+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.16517v2","created_at":"2026-07-01T01:17:53.017587+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.16517","created_at":"2026-07-01T01:17:53.017587+00:00"},{"alias_kind":"pith_short_12","alias_value":"6C7TQIUPWFIR","created_at":"2026-07-01T01:17:53.017587+00:00"},{"alias_kind":"pith_short_16","alias_value":"6C7TQIUPWFIR2VT7","created_at":"2026-07-01T01:17:53.017587+00:00"},{"alias_kind":"pith_short_8","alias_value":"6C7TQIUP","created_at":"2026-07-01T01:17:53.017587+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/6C7TQIUPWFIR2VT7WWK73RKDDT","json":"https://pith.science/pith/6C7TQIUPWFIR2VT7WWK73RKDDT.json","graph_json":"https://pith.science/api/pith-number/6C7TQIUPWFIR2VT7WWK73RKDDT/graph.json","events_json":"https://pith.science/api/pith-number/6C7TQIUPWFIR2VT7WWK73RKDDT/events.json","paper":"https://pith.science/paper/6C7TQIUP"},"agent_actions":{"view_html":"https://pith.science/pith/6C7TQIUPWFIR2VT7WWK73RKDDT","download_json":"https://pith.science/pith/6C7TQIUPWFIR2VT7WWK73RKDDT.json","view_paper":"https://pith.science/paper/6C7TQIUP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.16517&json=true","fetch_graph":"https://pith.science/api/pith-number/6C7TQIUPWFIR2VT7WWK73RKDDT/graph.json","fetch_events":"https://pith.science/api/pith-number/6C7TQIUPWFIR2VT7WWK73RKDDT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6C7TQIUPWFIR2VT7WWK73RKDDT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6C7TQIUPWFIR2VT7WWK73RKDDT/action/storage_attestation","attest_author":"https://pith.science/pith/6C7TQIUPWFIR2VT7WWK73RKDDT/action/author_attestation","sign_citation":"https://pith.science/pith/6C7TQIUPWFIR2VT7WWK73RKDDT/action/citation_signature","submit_replication":"https://pith.science/pith/6C7TQIUPWFIR2VT7WWK73RKDDT/action/replication_record"}},"created_at":"2026-07-01T01:17:53.017587+00:00","updated_at":"2026-07-01T01:17:53.017587+00:00"}