{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:L5HOVRFZW6C5ZDSUHRBE2EJYZ6","short_pith_number":"pith:L5HOVRFZ","schema_version":"1.0","canonical_sha256":"5f4eeac4b9b785dc8e543c424d1138cf8fef39e5d215bb51ea7e85981fdb2e12","source":{"kind":"arxiv","id":"2501.01702","version":2},"attestation_state":"computed","paper":{"title":"AgentRefine: Enhancing Agent Generalization through Refinement Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.RO"],"primary_cat":"cs.AI","authors_text":"Dayuan Fu, Jingang Wang, Keqing He, Weihao Zeng, Weiran Xu, Wei Wang, Wentao Hong, Xunliang Cai, Yejie Wang, Zhuoma Gongque","submitted_at":"2025-01-03T08:55:19Z","abstract_excerpt":"Large Language Model (LLM) based agents have proved their ability to perform complex tasks like humans. However, there is still a large gap between open-sourced LLMs and commercial models like the GPT series. In this paper, we focus on improving the agent generalization capabilities of LLMs via instruction tuning. We first observe that the existing agent training corpus exhibits satisfactory results on held-in evaluation sets but fails to generalize to held-out sets. These agent-tuning works face severe formatting errors and are frequently stuck in the same mistake for a long while. We analyze"},"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":"2501.01702","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-01-03T08:55:19Z","cross_cats_sorted":["cs.CL","cs.RO"],"title_canon_sha256":"9ac9a863db3a4d5649a94b33c63f69bb01571509e5eb12731605ea0bcedc38cb","abstract_canon_sha256":"f16d804337435628e607e35ceeba576c08ba6975ff4aa449547c54751a111218"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:18:47.803218Z","signature_b64":"qld3KxRCcsuQ6xyNBYbKJ5AwQ6dBNSmeynGZ+VS4T7FKBxZthRgNLsHo+0ETZqQpC6LeOEDlmTkPfYh4ooZBCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5f4eeac4b9b785dc8e543c424d1138cf8fef39e5d215bb51ea7e85981fdb2e12","last_reissued_at":"2026-07-05T10:18:47.802659Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:18:47.802659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AgentRefine: Enhancing Agent Generalization through Refinement Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.RO"],"primary_cat":"cs.AI","authors_text":"Dayuan Fu, Jingang Wang, Keqing He, Weihao Zeng, Weiran Xu, Wei Wang, Wentao Hong, Xunliang Cai, Yejie Wang, Zhuoma Gongque","submitted_at":"2025-01-03T08:55:19Z","abstract_excerpt":"Large Language Model (LLM) based agents have proved their ability to perform complex tasks like humans. However, there is still a large gap between open-sourced LLMs and commercial models like the GPT series. In this paper, we focus on improving the agent generalization capabilities of LLMs via instruction tuning. We first observe that the existing agent training corpus exhibits satisfactory results on held-in evaluation sets but fails to generalize to held-out sets. These agent-tuning works face severe formatting errors and are frequently stuck in the same mistake for a long while. We analyze"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.01702","kind":"arxiv","version":2},"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/2501.01702/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":"2501.01702","created_at":"2026-07-05T10:18:47.802729+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.01702v2","created_at":"2026-07-05T10:18:47.802729+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.01702","created_at":"2026-07-05T10:18:47.802729+00:00"},{"alias_kind":"pith_short_12","alias_value":"L5HOVRFZW6C5","created_at":"2026-07-05T10:18:47.802729+00:00"},{"alias_kind":"pith_short_16","alias_value":"L5HOVRFZW6C5ZDSU","created_at":"2026-07-05T10:18:47.802729+00:00"},{"alias_kind":"pith_short_8","alias_value":"L5HOVRFZ","created_at":"2026-07-05T10:18:47.802729+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17628","citing_title":"OPD-Evolver: Cultivating Holistic Agent Evolver via On-Policy Distillation","ref_index":118,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24828","citing_title":"Test-Time Deep Thinking to Explore Implicit Rules","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05413","citing_title":"From History to State: Constant-Context Skill Learning for LLM Agents","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6","json":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6.json","graph_json":"https://pith.science/api/pith-number/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/graph.json","events_json":"https://pith.science/api/pith-number/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/events.json","paper":"https://pith.science/paper/L5HOVRFZ"},"agent_actions":{"view_html":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6","download_json":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6.json","view_paper":"https://pith.science/paper/L5HOVRFZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.01702&json=true","fetch_graph":"https://pith.science/api/pith-number/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/graph.json","fetch_events":"https://pith.science/api/pith-number/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/action/storage_attestation","attest_author":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/action/author_attestation","sign_citation":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/action/citation_signature","submit_replication":"https://pith.science/pith/L5HOVRFZW6C5ZDSUHRBE2EJYZ6/action/replication_record"}},"created_at":"2026-07-05T10:18:47.802729+00:00","updated_at":"2026-07-05T10:18:47.802729+00:00"}