{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MUR5YBFGGEJ263DKKODENKCC3T","short_pith_number":"pith:MUR5YBFG","schema_version":"1.0","canonical_sha256":"6523dc04a63113af6c6a538646a842dcf3dcf828042485ca66a80ebad436304b","source":{"kind":"arxiv","id":"2605.31433","version":1},"attestation_state":"computed","paper":{"title":"SCOPE: Self-Play via Co-Evolving Policies for Open-Ended Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aryo Pradipta Gema, Joshua Ong Jun Leang, Pasquale Minervini, Wai-Chung Kwan","submitted_at":"2026-05-29T15:28:26Z","abstract_excerpt":"Self-play can train language models without external supervision. However, existing methods require rule-checkable answers, leaving open-ended tasks dependent on curated prompts or frontier-model judges. We introduce SCOPE, a data-free self-play framework for open-ended tasks that co-evolves two policies: a Challenger that generates document-grounded tasks, and a Solver that answers them through multi-turn retrieval. A frozen copy of the initial model serves as the self-judge, which writes task-specific rubrics from the source document and grades Solver responses against them. Across three 7-8"},"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":"2605.31433","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-29T15:28:26Z","cross_cats_sorted":[],"title_canon_sha256":"620b969b1447af61908e2853a9ac579147c3a9ef84c8f77c1261cadbc21270fb","abstract_canon_sha256":"acad12f954e229c95d413ef978afc99e4c613370e367893e755d1e7c9cc59c3d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T02:04:05.507590Z","signature_b64":"I9EO9XurZNUZLQ/AoPkq6pP3FHDVWjOWUbDwcvQXyKQYtvkje/BscJgiWh4xOEPTkyoifRQzBP1BiRKXzWO9Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6523dc04a63113af6c6a538646a842dcf3dcf828042485ca66a80ebad436304b","last_reissued_at":"2026-06-01T02:04:05.506645Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T02:04:05.506645Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SCOPE: Self-Play via Co-Evolving Policies for Open-Ended Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aryo Pradipta Gema, Joshua Ong Jun Leang, Pasquale Minervini, Wai-Chung Kwan","submitted_at":"2026-05-29T15:28:26Z","abstract_excerpt":"Self-play can train language models without external supervision. However, existing methods require rule-checkable answers, leaving open-ended tasks dependent on curated prompts or frontier-model judges. We introduce SCOPE, a data-free self-play framework for open-ended tasks that co-evolves two policies: a Challenger that generates document-grounded tasks, and a Solver that answers them through multi-turn retrieval. A frozen copy of the initial model serves as the self-judge, which writes task-specific rubrics from the source document and grades Solver responses against them. Across three 7-8"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.31433","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/2605.31433/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":"2605.31433","created_at":"2026-06-01T02:04:05.506747+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.31433v1","created_at":"2026-06-01T02:04:05.506747+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.31433","created_at":"2026-06-01T02:04:05.506747+00:00"},{"alias_kind":"pith_short_12","alias_value":"MUR5YBFGGEJ2","created_at":"2026-06-01T02:04:05.506747+00:00"},{"alias_kind":"pith_short_16","alias_value":"MUR5YBFGGEJ263DK","created_at":"2026-06-01T02:04:05.506747+00:00"},{"alias_kind":"pith_short_8","alias_value":"MUR5YBFG","created_at":"2026-06-01T02:04:05.506747+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/MUR5YBFGGEJ263DKKODENKCC3T","json":"https://pith.science/pith/MUR5YBFGGEJ263DKKODENKCC3T.json","graph_json":"https://pith.science/api/pith-number/MUR5YBFGGEJ263DKKODENKCC3T/graph.json","events_json":"https://pith.science/api/pith-number/MUR5YBFGGEJ263DKKODENKCC3T/events.json","paper":"https://pith.science/paper/MUR5YBFG"},"agent_actions":{"view_html":"https://pith.science/pith/MUR5YBFGGEJ263DKKODENKCC3T","download_json":"https://pith.science/pith/MUR5YBFGGEJ263DKKODENKCC3T.json","view_paper":"https://pith.science/paper/MUR5YBFG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.31433&json=true","fetch_graph":"https://pith.science/api/pith-number/MUR5YBFGGEJ263DKKODENKCC3T/graph.json","fetch_events":"https://pith.science/api/pith-number/MUR5YBFGGEJ263DKKODENKCC3T/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MUR5YBFGGEJ263DKKODENKCC3T/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MUR5YBFGGEJ263DKKODENKCC3T/action/storage_attestation","attest_author":"https://pith.science/pith/MUR5YBFGGEJ263DKKODENKCC3T/action/author_attestation","sign_citation":"https://pith.science/pith/MUR5YBFGGEJ263DKKODENKCC3T/action/citation_signature","submit_replication":"https://pith.science/pith/MUR5YBFGGEJ263DKKODENKCC3T/action/replication_record"}},"created_at":"2026-06-01T02:04:05.506747+00:00","updated_at":"2026-06-01T02:04:05.506747+00:00"}