{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YKFGCAL3JHJVT7XJOUKGJ7OPF3","short_pith_number":"pith:YKFGCAL3","schema_version":"1.0","canonical_sha256":"c28a61017b49d359fee9751464fdcf2eca2f13ff43c234a3ae72fd8c4116d7c5","source":{"kind":"arxiv","id":"2509.16456","version":3},"attestation_state":"computed","paper":{"title":"GPO: Learning from Critical Steps to Improve LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiahao Yu, Xian Wu, Xinyu Xing, Zelei Cheng","submitted_at":"2025-09-19T22:30:23Z","abstract_excerpt":"Large language models (LLMs) are increasingly used in various domains, showing impressive potential on different tasks. Recently, reasoning LLMs have been proposed to improve the \\textit{reasoning} or \\textit{thinking} capabilities of LLMs to solve complex problems. Despite the promising results of reasoning LLMs, enhancing the multi-step reasoning capabilities of LLMs still remains a significant challenge. While existing optimization methods have advanced the LLM reasoning capabilities, they often treat reasoning trajectories as a whole, without considering the underlying critical steps withi"},"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":"2509.16456","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-09-19T22:30:23Z","cross_cats_sorted":[],"title_canon_sha256":"db910f1808e80acd1486f1b72352425005f062a41fa0f84974a81ee9e20a0248","abstract_canon_sha256":"08e634b3504e36b02ed3569be51f67ddb118436204ba0b0b8b3129afa08d6b56"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-11T01:09:17.474389Z","signature_b64":"ydolK3w5n2ySNk8zyCTcseCQZATnNU0ix7xoHp2CkH9YKOZCsiO5LiAnengnCUc9D1tUHk48R49OpanEfCp+Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c28a61017b49d359fee9751464fdcf2eca2f13ff43c234a3ae72fd8c4116d7c5","last_reissued_at":"2026-06-11T01:09:17.473090Z","signature_status":"signed_v1","first_computed_at":"2026-06-11T01:09:17.473090Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"GPO: Learning from Critical Steps to Improve LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiahao Yu, Xian Wu, Xinyu Xing, Zelei Cheng","submitted_at":"2025-09-19T22:30:23Z","abstract_excerpt":"Large language models (LLMs) are increasingly used in various domains, showing impressive potential on different tasks. Recently, reasoning LLMs have been proposed to improve the \\textit{reasoning} or \\textit{thinking} capabilities of LLMs to solve complex problems. Despite the promising results of reasoning LLMs, enhancing the multi-step reasoning capabilities of LLMs still remains a significant challenge. While existing optimization methods have advanced the LLM reasoning capabilities, they often treat reasoning trajectories as a whole, without considering the underlying critical steps withi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.16456","kind":"arxiv","version":3},"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/2509.16456/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":"2509.16456","created_at":"2026-06-11T01:09:17.473261+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.16456v3","created_at":"2026-06-11T01:09:17.473261+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.16456","created_at":"2026-06-11T01:09:17.473261+00:00"},{"alias_kind":"pith_short_12","alias_value":"YKFGCAL3JHJV","created_at":"2026-06-11T01:09:17.473261+00:00"},{"alias_kind":"pith_short_16","alias_value":"YKFGCAL3JHJVT7XJ","created_at":"2026-06-11T01:09:17.473261+00:00"},{"alias_kind":"pith_short_8","alias_value":"YKFGCAL3","created_at":"2026-06-11T01:09:17.473261+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.03647","citing_title":"Stabilizing Unsupervised Self-Evolution of MLLMs via Continuous Softened Retracing reSampling","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3","json":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3.json","graph_json":"https://pith.science/api/pith-number/YKFGCAL3JHJVT7XJOUKGJ7OPF3/graph.json","events_json":"https://pith.science/api/pith-number/YKFGCAL3JHJVT7XJOUKGJ7OPF3/events.json","paper":"https://pith.science/paper/YKFGCAL3"},"agent_actions":{"view_html":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3","download_json":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3.json","view_paper":"https://pith.science/paper/YKFGCAL3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.16456&json=true","fetch_graph":"https://pith.science/api/pith-number/YKFGCAL3JHJVT7XJOUKGJ7OPF3/graph.json","fetch_events":"https://pith.science/api/pith-number/YKFGCAL3JHJVT7XJOUKGJ7OPF3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3/action/storage_attestation","attest_author":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3/action/author_attestation","sign_citation":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3/action/citation_signature","submit_replication":"https://pith.science/pith/YKFGCAL3JHJVT7XJOUKGJ7OPF3/action/replication_record"}},"created_at":"2026-06-11T01:09:17.473261+00:00","updated_at":"2026-06-11T01:09:17.473261+00:00"}