{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:VA36OPPEJDCBYXWDYU4KAJ5JS6","short_pith_number":"pith:VA36OPPE","schema_version":"1.0","canonical_sha256":"a837e73de448c41c5ec3c538a027a9979014db3a615c711c0315e2bab1d58d83","source":{"kind":"arxiv","id":"2607.13988","version":1},"attestation_state":"computed","paper":{"title":"TRACE: Turn-level Reward Assignment via Credit Estimation for Long-Horizon Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Baolin Peng, Hao Cheng, Jianfeng Gao, Leitian Tao, Mike Hang Wang, Sharon Li, Tao Ge, Wenlin Yao","submitted_at":"2026-07-15T16:16:42Z","abstract_excerpt":"Multi-turn agents solve complex tasks through extended sequences of tool interactions before producing a final answer, making credit assignment a fundamental challenge during post-training. Outcome rewards provide reliable supervision for short-horizon reasoning, but become sparse and high-variance as trajectories grow to tens or hundreds of tool calls. They can also be misleading: a failed rollout may contain many useful actions that move the agent closer to the goal, yet outcome-only training assigns them the same negative advantage as the eventual mistake. We propose TRACE (Turn-level Rewar"},"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":"2607.13988","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-15T16:16:42Z","cross_cats_sorted":[],"title_canon_sha256":"fd50588270e247f06840d4f0f9bde331ed76afaf117fe21405763705db40e6a0","abstract_canon_sha256":"022c21ca4f29b9412e5ce4be2168a65119cc840abbbb58203bc07bb0bf54346b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-16T01:23:18.168477Z","signature_b64":"46sKotuxSPcPeCpPl1d01Ix/izxitvKhk0G1TbMumgBVt/sLxIeGxjklG/qz2sQj6cMgwjcDdg4pIccMawAZDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a837e73de448c41c5ec3c538a027a9979014db3a615c711c0315e2bab1d58d83","last_reissued_at":"2026-07-16T01:23:18.167639Z","signature_status":"signed_v1","first_computed_at":"2026-07-16T01:23:18.167639Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TRACE: Turn-level Reward Assignment via Credit Estimation for Long-Horizon Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Baolin Peng, Hao Cheng, Jianfeng Gao, Leitian Tao, Mike Hang Wang, Sharon Li, Tao Ge, Wenlin Yao","submitted_at":"2026-07-15T16:16:42Z","abstract_excerpt":"Multi-turn agents solve complex tasks through extended sequences of tool interactions before producing a final answer, making credit assignment a fundamental challenge during post-training. Outcome rewards provide reliable supervision for short-horizon reasoning, but become sparse and high-variance as trajectories grow to tens or hundreds of tool calls. They can also be misleading: a failed rollout may contain many useful actions that move the agent closer to the goal, yet outcome-only training assigns them the same negative advantage as the eventual mistake. We propose TRACE (Turn-level Rewar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.13988","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/2607.13988/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":"2607.13988","created_at":"2026-07-16T01:23:18.168066+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.13988v1","created_at":"2026-07-16T01:23:18.168066+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.13988","created_at":"2026-07-16T01:23:18.168066+00:00"},{"alias_kind":"pith_short_12","alias_value":"VA36OPPEJDCB","created_at":"2026-07-16T01:23:18.168066+00:00"},{"alias_kind":"pith_short_16","alias_value":"VA36OPPEJDCBYXWD","created_at":"2026-07-16T01:23:18.168066+00:00"},{"alias_kind":"pith_short_8","alias_value":"VA36OPPE","created_at":"2026-07-16T01:23:18.168066+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/VA36OPPEJDCBYXWDYU4KAJ5JS6","json":"https://pith.science/pith/VA36OPPEJDCBYXWDYU4KAJ5JS6.json","graph_json":"https://pith.science/api/pith-number/VA36OPPEJDCBYXWDYU4KAJ5JS6/graph.json","events_json":"https://pith.science/api/pith-number/VA36OPPEJDCBYXWDYU4KAJ5JS6/events.json","paper":"https://pith.science/paper/VA36OPPE"},"agent_actions":{"view_html":"https://pith.science/pith/VA36OPPEJDCBYXWDYU4KAJ5JS6","download_json":"https://pith.science/pith/VA36OPPEJDCBYXWDYU4KAJ5JS6.json","view_paper":"https://pith.science/paper/VA36OPPE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.13988&json=true","fetch_graph":"https://pith.science/api/pith-number/VA36OPPEJDCBYXWDYU4KAJ5JS6/graph.json","fetch_events":"https://pith.science/api/pith-number/VA36OPPEJDCBYXWDYU4KAJ5JS6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VA36OPPEJDCBYXWDYU4KAJ5JS6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VA36OPPEJDCBYXWDYU4KAJ5JS6/action/storage_attestation","attest_author":"https://pith.science/pith/VA36OPPEJDCBYXWDYU4KAJ5JS6/action/author_attestation","sign_citation":"https://pith.science/pith/VA36OPPEJDCBYXWDYU4KAJ5JS6/action/citation_signature","submit_replication":"https://pith.science/pith/VA36OPPEJDCBYXWDYU4KAJ5JS6/action/replication_record"}},"created_at":"2026-07-16T01:23:18.168066+00:00","updated_at":"2026-07-16T01:23:18.168066+00:00"}