{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:O73PNWWMAFQAWN4NVP2IZHYDGD","short_pith_number":"pith:O73PNWWM","schema_version":"1.0","canonical_sha256":"77f6f6dacc01600b378dabf48c9f0330fbde013ae6209795fddc9c1216d6f7b2","source":{"kind":"arxiv","id":"2605.21996","version":1},"attestation_state":"computed","paper":{"title":"From Patches to Trajectories: Privileged Process Supervision for Software-Engineering Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Jin Song Dong, Murong Ma, Peng Cheng, Qinglin Zhu, Shuai Lu, Tianyu Chen, Yan Lu, Yeyun Gong, Yun Lin, Zhiyong Huang","submitted_at":"2026-05-21T04:54:55Z","abstract_excerpt":"Supervised fine-tuning (SFT) on long teacher trajectories is the dominant way to instill investigation and reasoning in open software-engineering (SWE) agents. Since every retained response becomes an imitation target, the student inherits the final outcome and intermediate flaws, including ungrounded leaps and redundant loops. High-quality training data must be effective(each step is grounded and narrows the agent's epistemic gap to the correct fix) and efficient(each step is information-bearing rather than redundant or looping). Existing recipes filter or relabel teacher rollouts using only "},"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":"2605.21996","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-05-21T04:54:55Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"86341a87a9d24197afebad8605d075ada98fbef5b0dbed962fdf7a2870961a44","abstract_canon_sha256":"b52ffea40b5804d6912c8b6348498d5b47ed7d3244f39453287c22b532f67f45"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-22T01:04:19.677383Z","signature_b64":"SjF/fvkXe6iZJvZyTqGj1yBXRqJeIW3mnqLf/gXltmmKa3OnJ4yK94fzkBHAoQLP0wba6xF6Y0Htf76o88dZBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"77f6f6dacc01600b378dabf48c9f0330fbde013ae6209795fddc9c1216d6f7b2","last_reissued_at":"2026-05-22T01:04:19.676681Z","signature_status":"signed_v1","first_computed_at":"2026-05-22T01:04:19.676681Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Patches to Trajectories: Privileged Process Supervision for Software-Engineering Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Jin Song Dong, Murong Ma, Peng Cheng, Qinglin Zhu, Shuai Lu, Tianyu Chen, Yan Lu, Yeyun Gong, Yun Lin, Zhiyong Huang","submitted_at":"2026-05-21T04:54:55Z","abstract_excerpt":"Supervised fine-tuning (SFT) on long teacher trajectories is the dominant way to instill investigation and reasoning in open software-engineering (SWE) agents. Since every retained response becomes an imitation target, the student inherits the final outcome and intermediate flaws, including ungrounded leaps and redundant loops. High-quality training data must be effective(each step is grounded and narrows the agent's epistemic gap to the correct fix) and efficient(each step is information-bearing rather than redundant or looping). Existing recipes filter or relabel teacher rollouts using only "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21996","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/2605.21996/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":"2605.21996","created_at":"2026-05-22T01:04:19.676812+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.21996v1","created_at":"2026-05-22T01:04:19.676812+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21996","created_at":"2026-05-22T01:04:19.676812+00:00"},{"alias_kind":"pith_short_12","alias_value":"O73PNWWMAFQA","created_at":"2026-05-22T01:04:19.676812+00:00"},{"alias_kind":"pith_short_16","alias_value":"O73PNWWMAFQAWN4N","created_at":"2026-05-22T01:04:19.676812+00:00"},{"alias_kind":"pith_short_8","alias_value":"O73PNWWM","created_at":"2026-05-22T01:04:19.676812+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/O73PNWWMAFQAWN4NVP2IZHYDGD","json":"https://pith.science/pith/O73PNWWMAFQAWN4NVP2IZHYDGD.json","graph_json":"https://pith.science/api/pith-number/O73PNWWMAFQAWN4NVP2IZHYDGD/graph.json","events_json":"https://pith.science/api/pith-number/O73PNWWMAFQAWN4NVP2IZHYDGD/events.json","paper":"https://pith.science/paper/O73PNWWM"},"agent_actions":{"view_html":"https://pith.science/pith/O73PNWWMAFQAWN4NVP2IZHYDGD","download_json":"https://pith.science/pith/O73PNWWMAFQAWN4NVP2IZHYDGD.json","view_paper":"https://pith.science/paper/O73PNWWM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.21996&json=true","fetch_graph":"https://pith.science/api/pith-number/O73PNWWMAFQAWN4NVP2IZHYDGD/graph.json","fetch_events":"https://pith.science/api/pith-number/O73PNWWMAFQAWN4NVP2IZHYDGD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O73PNWWMAFQAWN4NVP2IZHYDGD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O73PNWWMAFQAWN4NVP2IZHYDGD/action/storage_attestation","attest_author":"https://pith.science/pith/O73PNWWMAFQAWN4NVP2IZHYDGD/action/author_attestation","sign_citation":"https://pith.science/pith/O73PNWWMAFQAWN4NVP2IZHYDGD/action/citation_signature","submit_replication":"https://pith.science/pith/O73PNWWMAFQAWN4NVP2IZHYDGD/action/replication_record"}},"created_at":"2026-05-22T01:04:19.676812+00:00","updated_at":"2026-05-22T01:04:19.676812+00:00"}