{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:6ZPD3FOASM6DTN43Y3BKLRUTEH","short_pith_number":"pith:6ZPD3FOA","schema_version":"1.0","canonical_sha256":"f65e3d95c0933c39b79bc6c2a5c69321fab0c14345cce071fc07cfd2e37c6529","source":{"kind":"arxiv","id":"2403.18421","version":1},"attestation_state":"computed","paper":{"title":"BioMedLM: A 2.7B Parameter Language Model Trained On Biomedical Text","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Abhinav Venigalla, Betty Xiong, Christopher D. Manning, David Hall, Elliot Bolton, Jonathan Frankle, Michael Carbin, Michihiro Yasunaga, Percy Liang, Roxana Daneshjou, Tony Lee","submitted_at":"2024-03-27T10:18:21Z","abstract_excerpt":"Models such as GPT-4 and Med-PaLM 2 have demonstrated impressive performance on a wide variety of biomedical NLP tasks. However, these models have hundreds of billions of parameters, are computationally expensive to run, require users to send their input data over the internet, and are trained on unknown data sources. Can smaller, more targeted models compete? To address this question, we build and release BioMedLM, a 2.7 billion parameter GPT-style autoregressive model trained exclusively on PubMed abstracts and full articles. When fine-tuned, BioMedLM can produce strong multiple-choice biome"},"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":"2403.18421","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-03-27T10:18:21Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ba4dc9a1d621a96f84dade2d19f3e4c02bd59f7d33dc98d7d2d9146c6227bd27","abstract_canon_sha256":"a0cab908cca13a4247c13527617fc5de5ec1e043e3b5db11c40ffef89b6731e4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:01:22.432930Z","signature_b64":"6+8jcsOKlld7L1AbPEkhAVCg/2uoqroBDF32UuiKw/FTlvGzHwcQPNQZ6+E13rkPLOcHp6Wte9ysv8Y1ncrBDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f65e3d95c0933c39b79bc6c2a5c69321fab0c14345cce071fc07cfd2e37c6529","last_reissued_at":"2026-07-05T08:01:22.432347Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:01:22.432347Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"BioMedLM: A 2.7B Parameter Language Model Trained On Biomedical Text","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Abhinav Venigalla, Betty Xiong, Christopher D. Manning, David Hall, Elliot Bolton, Jonathan Frankle, Michael Carbin, Michihiro Yasunaga, Percy Liang, Roxana Daneshjou, Tony Lee","submitted_at":"2024-03-27T10:18:21Z","abstract_excerpt":"Models such as GPT-4 and Med-PaLM 2 have demonstrated impressive performance on a wide variety of biomedical NLP tasks. However, these models have hundreds of billions of parameters, are computationally expensive to run, require users to send their input data over the internet, and are trained on unknown data sources. Can smaller, more targeted models compete? To address this question, we build and release BioMedLM, a 2.7 billion parameter GPT-style autoregressive model trained exclusively on PubMed abstracts and full articles. When fine-tuned, BioMedLM can produce strong multiple-choice biome"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.18421","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/2403.18421/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":"2403.18421","created_at":"2026-07-05T08:01:22.432444+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.18421v1","created_at":"2026-07-05T08:01:22.432444+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.18421","created_at":"2026-07-05T08:01:22.432444+00:00"},{"alias_kind":"pith_short_12","alias_value":"6ZPD3FOASM6D","created_at":"2026-07-05T08:01:22.432444+00:00"},{"alias_kind":"pith_short_16","alias_value":"6ZPD3FOASM6DTN43","created_at":"2026-07-05T08:01:22.432444+00:00"},{"alias_kind":"pith_short_8","alias_value":"6ZPD3FOA","created_at":"2026-07-05T08:01:22.432444+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17213","citing_title":"Revisiting LLM Adaptation for 3D CT Report Generation: A Study of Scaling and Diagnostic Priors","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2503.08223","citing_title":"Will LLMs Scaling Hit the Wall? Breaking Barriers via Distributed Resources on Massive Edge Devices","ref_index":93,"is_internal_anchor":false},{"citing_arxiv_id":"2509.07177","citing_title":"Towards EnergyGPT: A Large Language Model Specialized for the Energy Sector","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11774","citing_title":"From Token to Token Pair: Efficient Prompt Compression for Large Language Models in Clinical Prediction","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH","json":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH.json","graph_json":"https://pith.science/api/pith-number/6ZPD3FOASM6DTN43Y3BKLRUTEH/graph.json","events_json":"https://pith.science/api/pith-number/6ZPD3FOASM6DTN43Y3BKLRUTEH/events.json","paper":"https://pith.science/paper/6ZPD3FOA"},"agent_actions":{"view_html":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH","download_json":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH.json","view_paper":"https://pith.science/paper/6ZPD3FOA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.18421&json=true","fetch_graph":"https://pith.science/api/pith-number/6ZPD3FOASM6DTN43Y3BKLRUTEH/graph.json","fetch_events":"https://pith.science/api/pith-number/6ZPD3FOASM6DTN43Y3BKLRUTEH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH/action/storage_attestation","attest_author":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH/action/author_attestation","sign_citation":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH/action/citation_signature","submit_replication":"https://pith.science/pith/6ZPD3FOASM6DTN43Y3BKLRUTEH/action/replication_record"}},"created_at":"2026-07-05T08:01:22.432444+00:00","updated_at":"2026-07-05T08:01:22.432444+00:00"}