{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3VFE4BAQR44C2NAS4NEX22CYUN","short_pith_number":"pith:3VFE4BAQ","schema_version":"1.0","canonical_sha256":"dd4a4e04108f382d3412e3497d6858a3752c3096251456e9f8a0646fcb30836f","source":{"kind":"arxiv","id":"2607.00531","version":1},"attestation_state":"computed","paper":{"title":"Active-GRPO: Adaptive Imitation and Self-Improving Reasoning for Molecular Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","q-bio.BM","stat.ML"],"primary_cat":"cs.LG","authors_text":"Le Cong, Mingxuan Cao, Qinan Huang, Rick Stevens, Thomas Brettin, Xuefeng Liu","submitted_at":"2026-07-01T07:22:46Z","abstract_excerpt":"Scientific reasoning is an increasingly important capability of large language models, yet improving the robustness and efficiency of training such reasoning remains a key open challenge. We study this problem in instruction-based molecular optimization, where answer-only supervised fine-tuning (SFT) collapses multi-step reasoning and reinforcement learning with verifiable rewards (RLVR) suffers from sparse feedback. Reference-guided Policy Optimization mitigates both by anchoring policy updates to dataset-provided references, but its effectiveness is tightly coupled to reference quality: weak"},"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":"2607.00531","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-07-01T07:22:46Z","cross_cats_sorted":["cs.AI","q-bio.BM","stat.ML"],"title_canon_sha256":"b00faf0485e4a6044f5e0c1a05da187489194672cfb3181d702fb23abe7b00a3","abstract_canon_sha256":"b9ee896275e9a119769a8d06a198d5daad47665c9c3db109600780b19b52d220"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-02T01:17:46.783103Z","signature_b64":"VXtxiL45A7/WUwoJidijZOdQbifGoFXn9qQwWzmY3HwMa3UmXTDxvonDLdJuviysAeD2dKLn3VaI5MqeWKBNDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dd4a4e04108f382d3412e3497d6858a3752c3096251456e9f8a0646fcb30836f","last_reissued_at":"2026-07-02T01:17:46.782699Z","signature_status":"signed_v1","first_computed_at":"2026-07-02T01:17:46.782699Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Active-GRPO: Adaptive Imitation and Self-Improving Reasoning for Molecular Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","q-bio.BM","stat.ML"],"primary_cat":"cs.LG","authors_text":"Le Cong, Mingxuan Cao, Qinan Huang, Rick Stevens, Thomas Brettin, Xuefeng Liu","submitted_at":"2026-07-01T07:22:46Z","abstract_excerpt":"Scientific reasoning is an increasingly important capability of large language models, yet improving the robustness and efficiency of training such reasoning remains a key open challenge. We study this problem in instruction-based molecular optimization, where answer-only supervised fine-tuning (SFT) collapses multi-step reasoning and reinforcement learning with verifiable rewards (RLVR) suffers from sparse feedback. Reference-guided Policy Optimization mitigates both by anchoring policy updates to dataset-provided references, but its effectiveness is tightly coupled to reference quality: weak"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.00531","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/2607.00531/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":"2607.00531","created_at":"2026-07-02T01:17:46.782756+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.00531v1","created_at":"2026-07-02T01:17:46.782756+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.00531","created_at":"2026-07-02T01:17:46.782756+00:00"},{"alias_kind":"pith_short_12","alias_value":"3VFE4BAQR44C","created_at":"2026-07-02T01:17:46.782756+00:00"},{"alias_kind":"pith_short_16","alias_value":"3VFE4BAQR44C2NAS","created_at":"2026-07-02T01:17:46.782756+00:00"},{"alias_kind":"pith_short_8","alias_value":"3VFE4BAQ","created_at":"2026-07-02T01:17:46.782756+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/3VFE4BAQR44C2NAS4NEX22CYUN","json":"https://pith.science/pith/3VFE4BAQR44C2NAS4NEX22CYUN.json","graph_json":"https://pith.science/api/pith-number/3VFE4BAQR44C2NAS4NEX22CYUN/graph.json","events_json":"https://pith.science/api/pith-number/3VFE4BAQR44C2NAS4NEX22CYUN/events.json","paper":"https://pith.science/paper/3VFE4BAQ"},"agent_actions":{"view_html":"https://pith.science/pith/3VFE4BAQR44C2NAS4NEX22CYUN","download_json":"https://pith.science/pith/3VFE4BAQR44C2NAS4NEX22CYUN.json","view_paper":"https://pith.science/paper/3VFE4BAQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.00531&json=true","fetch_graph":"https://pith.science/api/pith-number/3VFE4BAQR44C2NAS4NEX22CYUN/graph.json","fetch_events":"https://pith.science/api/pith-number/3VFE4BAQR44C2NAS4NEX22CYUN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3VFE4BAQR44C2NAS4NEX22CYUN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3VFE4BAQR44C2NAS4NEX22CYUN/action/storage_attestation","attest_author":"https://pith.science/pith/3VFE4BAQR44C2NAS4NEX22CYUN/action/author_attestation","sign_citation":"https://pith.science/pith/3VFE4BAQR44C2NAS4NEX22CYUN/action/citation_signature","submit_replication":"https://pith.science/pith/3VFE4BAQR44C2NAS4NEX22CYUN/action/replication_record"}},"created_at":"2026-07-02T01:17:46.782756+00:00","updated_at":"2026-07-02T01:17:46.782756+00:00"}