{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JZJ3UO3LPPU3PSJIXED26KS5CQ","short_pith_number":"pith:JZJ3UO3L","schema_version":"1.0","canonical_sha256":"4e53ba3b6b7be9b7c928b907af2a5d142d47a07ccf3664a5b98b07b09a1d28b0","source":{"kind":"arxiv","id":"2606.26563","version":1},"attestation_state":"computed","paper":{"title":"scBench-Long: Verifiable Benchmarking of Long-Horizon Single-Cell Biology","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"q-bio.GN","authors_text":"Arjun Banerjee, Ian Diks, Kenny Workman, Tim Proctor, Zhen Yang","submitted_at":"2026-06-25T03:21:50Z","abstract_excerpt":"Single-cell studies require analysts to convert raw measurements into specific biological claims through multi-step workflows and integration of metadata, assay context, and auxiliary evidence. Existing AI-biology benchmarks largely measure broad knowledge, executable workflows, or local analysis steps. We introduce scBench-Long, a benchmark for long-horizon single-cell biology in which agents must recover scientific conclusions from raw or near-raw data without prescribed methods. The benchmark contains 21 evaluations spanning melanoma CD8 T-cell reactivity, CD8 RNA+ATAC regulatory inference,"},"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":"2606.26563","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"q-bio.GN","submitted_at":"2026-06-25T03:21:50Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"43750902a08b487349621f7c06228f0e9cdbc43b90946384bb772db90070056f","abstract_canon_sha256":"1c16525e80df9cffecbfd046a402aba5bae52bcaf4fcf9c0dbb649cf3b2ac6d3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-26T01:15:34.820508Z","signature_b64":"1fvM+6QuaiChgqAyWj5oh54ww27tht7alV9Ob3rWhrRsPFJ78lwOsDezLHCwvyENRsLndlQzTgLGlyGKIo/3Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4e53ba3b6b7be9b7c928b907af2a5d142d47a07ccf3664a5b98b07b09a1d28b0","last_reissued_at":"2026-06-26T01:15:34.820092Z","signature_status":"signed_v1","first_computed_at":"2026-06-26T01:15:34.820092Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"scBench-Long: Verifiable Benchmarking of Long-Horizon Single-Cell Biology","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"q-bio.GN","authors_text":"Arjun Banerjee, Ian Diks, Kenny Workman, Tim Proctor, Zhen Yang","submitted_at":"2026-06-25T03:21:50Z","abstract_excerpt":"Single-cell studies require analysts to convert raw measurements into specific biological claims through multi-step workflows and integration of metadata, assay context, and auxiliary evidence. Existing AI-biology benchmarks largely measure broad knowledge, executable workflows, or local analysis steps. We introduce scBench-Long, a benchmark for long-horizon single-cell biology in which agents must recover scientific conclusions from raw or near-raw data without prescribed methods. The benchmark contains 21 evaluations spanning melanoma CD8 T-cell reactivity, CD8 RNA+ATAC regulatory inference,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26563","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/2606.26563/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":"2606.26563","created_at":"2026-06-26T01:15:34.820153+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.26563v1","created_at":"2026-06-26T01:15:34.820153+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26563","created_at":"2026-06-26T01:15:34.820153+00:00"},{"alias_kind":"pith_short_12","alias_value":"JZJ3UO3LPPU3","created_at":"2026-06-26T01:15:34.820153+00:00"},{"alias_kind":"pith_short_16","alias_value":"JZJ3UO3LPPU3PSJI","created_at":"2026-06-26T01:15:34.820153+00:00"},{"alias_kind":"pith_short_8","alias_value":"JZJ3UO3L","created_at":"2026-06-26T01:15:34.820153+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/JZJ3UO3LPPU3PSJIXED26KS5CQ","json":"https://pith.science/pith/JZJ3UO3LPPU3PSJIXED26KS5CQ.json","graph_json":"https://pith.science/api/pith-number/JZJ3UO3LPPU3PSJIXED26KS5CQ/graph.json","events_json":"https://pith.science/api/pith-number/JZJ3UO3LPPU3PSJIXED26KS5CQ/events.json","paper":"https://pith.science/paper/JZJ3UO3L"},"agent_actions":{"view_html":"https://pith.science/pith/JZJ3UO3LPPU3PSJIXED26KS5CQ","download_json":"https://pith.science/pith/JZJ3UO3LPPU3PSJIXED26KS5CQ.json","view_paper":"https://pith.science/paper/JZJ3UO3L","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.26563&json=true","fetch_graph":"https://pith.science/api/pith-number/JZJ3UO3LPPU3PSJIXED26KS5CQ/graph.json","fetch_events":"https://pith.science/api/pith-number/JZJ3UO3LPPU3PSJIXED26KS5CQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JZJ3UO3LPPU3PSJIXED26KS5CQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JZJ3UO3LPPU3PSJIXED26KS5CQ/action/storage_attestation","attest_author":"https://pith.science/pith/JZJ3UO3LPPU3PSJIXED26KS5CQ/action/author_attestation","sign_citation":"https://pith.science/pith/JZJ3UO3LPPU3PSJIXED26KS5CQ/action/citation_signature","submit_replication":"https://pith.science/pith/JZJ3UO3LPPU3PSJIXED26KS5CQ/action/replication_record"}},"created_at":"2026-06-26T01:15:34.820153+00:00","updated_at":"2026-06-26T01:15:34.820153+00:00"}