{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2606-16517","title":"How Post-Training Shapes Biological Reasoning Models","arxiv_id":"2606.16517","date":"2026-06-15","proceeding":null,"authors":["Lukas Fesser","Hanlin Zhang","Michelle M. Li","Eric Wang","Bryan Perozzi","Shekoofeh Azizi","Sham M. Kakade","Marinka Zitnik"],"abstract":"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 reinforcement learning (RL), measuring both in-domain (ID) and out-of-domain (OOD) performance. We find that each post-training stage reshapes generalization in a distinct way rather than contributing uniform gains. CPT improves downstream performance by aligning models with biological language. SFT consistently increases ID performance but causes OOD performance to peak early and decline as models fit the training distribution. RL, when applied to strong SFT checkpoints with aligned rewards, improves OOD performance and partially recovers generalization. These results show that biological reasoning does not improve monotonically with additional supervision or compute. Instead, performance depends on how training stages are composed. Under fixed post-training budgets, the strongest ID-OOD trade-off comes from brief SFT, larger RL allocations, and asymmetric adaptation capacity across stages.","url_abs":"https://arxiv.org/abs/2606.16517","url_pdf":"https://arxiv.org/pdf/2606.16517","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2606.16517","atlas_url":"https://app.syntology.ai/?focus=2606.16517","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2606.16517"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/mims-harvard/bio-posttrain","reach":null}],"summary":{"ran":12,"unverified":4},"by_repo_kind":{"found_in_text":{"samples":16,"ran":12,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":16,"samples":[{"code_sha256_prefix":"3bb51e9b7e0198cd","entry":"best_rl","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason/gen_post_cpt_report.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason/gen_post_cpt_report.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"3bb51e9b7e0198cd"}},{"code_sha256_prefix":"22389671b84fc915","entry":"best_sft_acc","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason/gen_post_cpt_report.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason/gen_post_cpt_report.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"22389671b84fc915"}},{"code_sha256_prefix":"e15f79e855c0f54b","entry":"clean_sequence","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason-Pro/predict.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason-Pro/predict.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"e15f79e855c0f54b"}},{"code_sha256_prefix":"6d5992a248b4c807","entry":"extract_hash_answer","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason/reason.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason/reason.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"6d5992a248b4c807"}},{"code_sha256_prefix":"3fca4c3b8039e70a","entry":"extract_xml_answer","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason/reason.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason/reason.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"3fca4c3b8039e70a"}},{"code_sha256_prefix":"aa05e6f2ab6b4e9b","entry":"harvest_dna_sft","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"harvest_results.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/harvest_results.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"aa05e6f2ab6b4e9b"}},{"code_sha256_prefix":"f14d500e7014c5ce","entry":"last_rl_reward","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason/gen_post_cpt_report.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason/gen_post_cpt_report.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"f14d500e7014c5ce"}},{"code_sha256_prefix":"584eae0463aafa3a","entry":"latest_log","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"harvest_results.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/harvest_results.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"584eae0463aafa3a"}},{"code_sha256_prefix":"f2a2993957885ef5","entry":"left_pad_to_max","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason-Pro/train_protein_grpo.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason-Pro/train_protein_grpo.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"f2a2993957885ef5"}},{"code_sha256_prefix":"e5c601c4aa7a3488","entry":"maybe_get","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason-Pro/train_protein_grpo.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason-Pro/train_protein_grpo.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"e5c601c4aa7a3488"}},{"code_sha256_prefix":"6923a8e4df002554","entry":"tail_text","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"harvest_results.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/harvest_results.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"6923a8e4df002554"}},{"code_sha256_prefix":"6a0ade46b1a945ae","entry":"validate_organism","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason-Pro/predict.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason-Pro/predict.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"6a0ade46b1a945ae"}},{"code_sha256_prefix":"b5d9ee083ccde995","entry":"load_interpro_metadata","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason-Pro/interpro_api.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason-Pro/interpro_api.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"b5d9ee083ccde995"}},{"code_sha256_prefix":"d6bc4ce169f90660","entry":"read_input_tsv","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason-Pro/predict.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason-Pro/predict.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"d6bc4ce169f90660"}},{"code_sha256_prefix":"1d08348ea56f3fea","entry":"run_interproscan_local","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason-Pro/interpro_api.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason-Pro/interpro_api.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"1d08348ea56f3fea"}},{"code_sha256_prefix":"a2eb2097b0dfa743","entry":"run_interproscan_online","repo":"mims-harvard/bio-posttrain","repo_kind":"found_in_text","path":"BioReason-Pro/interpro_api.py","file_url":"https://github.com/mims-harvard/bio-posttrain/blob/HEAD/BioReason-Pro/interpro_api.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"a2eb2097b0dfa743"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":{"kind":"leaderboard_placements","source":"Syntology's leaderboard-shaped extractor over the paper's own arXiv-HTML tables: a model pointed at a cell, the number was read from that cell and checked against the board's metric, dataset, split and scale, and an independent check accepted the entry; not reviewed by the paper's authors or the archive's editors","extractor_model":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","verifier_model":null,"prompt_sha":"fa63d4bb9d755694","coverage":{"sentence":"Syntology has checked 6,885 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","papers_newer_than_archive":9623,"papers_checked":6885},"entries":[],"not_placed":{"boards":0,"rejected_by_independent_check":0,"refused_by_a_rule":0,"check_did_not_answer":0,"proposed_without_a_cell":0,"declined_by_site":0}}}