{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:MXQRVVEXJCGCJPMFUNJNCJLYPA","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"19bebb70abcc094017accc4ca61f1e4e87e7646cabe32d4880b5270d3512afd2","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-21T16:36:19Z","title_canon_sha256":"bf80f8176593950e78c33aa3a2347d8ef5c7690b2869462eeba80a1522c5064a"},"schema_version":"1.0","source":{"id":"2601.15158","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2601.15158","created_at":"2026-06-04T01:08:40Z"},{"alias_kind":"arxiv_version","alias_value":"2601.15158v4","created_at":"2026-06-04T01:08:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.15158","created_at":"2026-06-04T01:08:40Z"},{"alias_kind":"pith_short_12","alias_value":"MXQRVVEXJCGC","created_at":"2026-06-04T01:08:40Z"},{"alias_kind":"pith_short_16","alias_value":"MXQRVVEXJCGCJPMF","created_at":"2026-06-04T01:08:40Z"},{"alias_kind":"pith_short_8","alias_value":"MXQRVVEX","created_at":"2026-06-04T01:08:40Z"}],"graph_snapshots":[{"event_id":"sha256:ce2fa824f03efc557c40752ca24108d625aaedd58336503454c2d35931552273","target":"graph","created_at":"2026-06-04T01:08:40Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2601.15158/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Transformers trained via Reinforcement Learning (RL) with outcome-based supervision can spontaneously develop the ability to generate intermediate reasoning steps (Chain-of-Thought). Yet the mechanism by which sparse rewards drive policy gradient to discover such systematic reasoning remains poorly understood. We address this by analyzing the policy gradient dynamics of single-layer Transformers on a synthetic graph traversal task that cannot be solved without Chain-of-Thought but admits a simple iterative solution. We prove that despite training solely on final-answer correctness, policy grad","authors_text":"Nadav Cohen, Shahar Mendel, Yotam Alexander, Yuval Ran-Milo","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-21T16:36:19Z","title":"Outcome-Based RL Provably Leads Transformers to Reason, but Only With the Right Data"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.15158","kind":"arxiv","version":4},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:bb964d02213f1c8597d145d9af3047bb47647297ea04be39247a6e25544ff3d2","target":"record","created_at":"2026-06-04T01:08:40Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"19bebb70abcc094017accc4ca61f1e4e87e7646cabe32d4880b5270d3512afd2","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-21T16:36:19Z","title_canon_sha256":"bf80f8176593950e78c33aa3a2347d8ef5c7690b2869462eeba80a1522c5064a"},"schema_version":"1.0","source":{"id":"2601.15158","kind":"arxiv","version":4}},"canonical_sha256":"65e11ad497488c24bd85a352d12578782110e39f9949f52cac3178f8cbc9b70f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"65e11ad497488c24bd85a352d12578782110e39f9949f52cac3178f8cbc9b70f","first_computed_at":"2026-06-04T01:08:40.289270Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-04T01:08:40.289270Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"7R06nSD0ATeesl6/lmmHEdhWNq1FAo3BlgT2eTR85jT2c2c+QhMeKAcP1p1CMWr3vSkXxQazmbmDNxaz364MDA==","signature_status":"signed_v1","signed_at":"2026-06-04T01:08:40.290097Z","signed_message":"canonical_sha256_bytes"},"source_id":"2601.15158","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bb964d02213f1c8597d145d9af3047bb47647297ea04be39247a6e25544ff3d2","sha256:ce2fa824f03efc557c40752ca24108d625aaedd58336503454c2d35931552273"],"state_sha256":"77e9d3758a7017a9b79fc7716da9a491135b0376a1450ff41cd57e674503b12c"}