{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:24DYVYK2EG4N2V2F5Y55TCSKCW","short_pith_number":"pith:24DYVYK2","schema_version":"1.0","canonical_sha256":"d7078ae15a21b8dd5745ee3bd98a4a15af6ce99e4da1bbd6cd57ed131425bbd4","source":{"kind":"arxiv","id":"2511.21667","version":4},"attestation_state":"computed","paper":{"title":"Escaping the Verifier: Learning to Reason via Demonstrations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ivan Provilkov, Locke Cai, Max Ryabinin","submitted_at":"2025-11-26T18:42:52Z","abstract_excerpt":"Training Large Language Models (LLMs) to reason often relies on Reinforcement Learning (RL) with task-specific verifiers. However, many real-world reasoning-intensive tasks lack verifiers, despite offering abundant expert demonstrations that remain under-utilized for reasoning-focused training. We introduce RARO (Relativistic Adversarial Reasoning Optimization), which learns strong reasoning capabilities from expert demonstrations alone via Inverse Reinforcement Learning. RARO sets up an adversarial game between a policy and a relativistic critic: the policy learns to mimic expert answers, whi"},"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":"2511.21667","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-11-26T18:42:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"be28bb4727885b6009197b0f861d9872d1dad43c25333ac91ceb7f0a93f27334","abstract_canon_sha256":"02bb3829985d2c84f4c6359ecc4d8f148409c51f23f53a28db65aeae38b51dbc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:14:31.302389Z","signature_b64":"OfEHTaP37xCXn67EgJV6P0hbNUkFDx8ZkTDvoDbc9q74j7AiXnuK5cDgNQoW4gLmXnRgYls0gbQW4kvwBZX0Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d7078ae15a21b8dd5745ee3bd98a4a15af6ce99e4da1bbd6cd57ed131425bbd4","last_reissued_at":"2026-06-05T01:14:31.301593Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:14:31.301593Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Escaping the Verifier: Learning to Reason via Demonstrations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ivan Provilkov, Locke Cai, Max Ryabinin","submitted_at":"2025-11-26T18:42:52Z","abstract_excerpt":"Training Large Language Models (LLMs) to reason often relies on Reinforcement Learning (RL) with task-specific verifiers. However, many real-world reasoning-intensive tasks lack verifiers, despite offering abundant expert demonstrations that remain under-utilized for reasoning-focused training. We introduce RARO (Relativistic Adversarial Reasoning Optimization), which learns strong reasoning capabilities from expert demonstrations alone via Inverse Reinforcement Learning. RARO sets up an adversarial game between a policy and a relativistic critic: the policy learns to mimic expert answers, whi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2511.21667","kind":"arxiv","version":4},"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/2511.21667/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":"2511.21667","created_at":"2026-06-05T01:14:31.301719+00:00"},{"alias_kind":"arxiv_version","alias_value":"2511.21667v4","created_at":"2026-06-05T01:14:31.301719+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.21667","created_at":"2026-06-05T01:14:31.301719+00:00"},{"alias_kind":"pith_short_12","alias_value":"24DYVYK2EG4N","created_at":"2026-06-05T01:14:31.301719+00:00"},{"alias_kind":"pith_short_16","alias_value":"24DYVYK2EG4N2V2F","created_at":"2026-06-05T01:14:31.301719+00:00"},{"alias_kind":"pith_short_8","alias_value":"24DYVYK2","created_at":"2026-06-05T01:14:31.301719+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/24DYVYK2EG4N2V2F5Y55TCSKCW","json":"https://pith.science/pith/24DYVYK2EG4N2V2F5Y55TCSKCW.json","graph_json":"https://pith.science/api/pith-number/24DYVYK2EG4N2V2F5Y55TCSKCW/graph.json","events_json":"https://pith.science/api/pith-number/24DYVYK2EG4N2V2F5Y55TCSKCW/events.json","paper":"https://pith.science/paper/24DYVYK2"},"agent_actions":{"view_html":"https://pith.science/pith/24DYVYK2EG4N2V2F5Y55TCSKCW","download_json":"https://pith.science/pith/24DYVYK2EG4N2V2F5Y55TCSKCW.json","view_paper":"https://pith.science/paper/24DYVYK2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2511.21667&json=true","fetch_graph":"https://pith.science/api/pith-number/24DYVYK2EG4N2V2F5Y55TCSKCW/graph.json","fetch_events":"https://pith.science/api/pith-number/24DYVYK2EG4N2V2F5Y55TCSKCW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/24DYVYK2EG4N2V2F5Y55TCSKCW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/24DYVYK2EG4N2V2F5Y55TCSKCW/action/storage_attestation","attest_author":"https://pith.science/pith/24DYVYK2EG4N2V2F5Y55TCSKCW/action/author_attestation","sign_citation":"https://pith.science/pith/24DYVYK2EG4N2V2F5Y55TCSKCW/action/citation_signature","submit_replication":"https://pith.science/pith/24DYVYK2EG4N2V2F5Y55TCSKCW/action/replication_record"}},"created_at":"2026-06-05T01:14:31.301719+00:00","updated_at":"2026-06-05T01:14:31.301719+00:00"}