{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:UFV56ISUTJ4XGPAI6JWJ5IALBF","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":"98637eb7c5abf570d3c2dd304ccb19c2601163f71b40bfe96805452fe469ccc4","cross_cats_sorted":["cs.AI","cs.CV"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T20:43:10Z","title_canon_sha256":"425078031f40dcc5516ad94c3b3b2f2a5b5b83b4fe88606d9ecd122e52ffe3a1"},"schema_version":"1.0","source":{"id":"2605.29089","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.29089","created_at":"2026-05-29T01:05:17Z"},{"alias_kind":"arxiv_version","alias_value":"2605.29089v1","created_at":"2026-05-29T01:05:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29089","created_at":"2026-05-29T01:05:17Z"},{"alias_kind":"pith_short_12","alias_value":"UFV56ISUTJ4X","created_at":"2026-05-29T01:05:17Z"},{"alias_kind":"pith_short_16","alias_value":"UFV56ISUTJ4XGPAI","created_at":"2026-05-29T01:05:17Z"},{"alias_kind":"pith_short_8","alias_value":"UFV56ISU","created_at":"2026-05-29T01:05:17Z"}],"graph_snapshots":[{"event_id":"sha256:5f7300521988358717b052570af4e08c524d96c489144531fc3ac9895663e985","target":"graph","created_at":"2026-05-29T01:05:17Z","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/2605.29089/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent reinforcement learning (RL) post-training approaches primarily optimize the final output policy using sparse outcome-level rewards, while largely overlooking predictive signals encoded in intermediate representations. In this paper, we introduce a new paradigm called on-policy internal self-distillation and propose the OISD framework, which improves reasoning by transferring on-policy predictive signals from the final layer to intermediate representations. During rollout and Group Relative Policy Optimization (GRPO) optimization, the final layer acts as both the policy and a detached in","authors_text":"Darryl Cherian Jacob, Jindong Wang, Pan He, Xinyu Liu, Yang Zhou","cross_cats":["cs.AI","cs.CV"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T20:43:10Z","title":"OISD: On-Policy Internal Self-Distillation of Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29089","kind":"arxiv","version":1},"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":"sha256:3fc6d2b8e8e5db66d70bd3faaed10b95ddfc36705b2e849dae759a27b7dab3bf","target":"record","created_at":"2026-05-29T01:05:17Z","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":"98637eb7c5abf570d3c2dd304ccb19c2601163f71b40bfe96805452fe469ccc4","cross_cats_sorted":["cs.AI","cs.CV"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T20:43:10Z","title_canon_sha256":"425078031f40dcc5516ad94c3b3b2f2a5b5b83b4fe88606d9ecd122e52ffe3a1"},"schema_version":"1.0","source":{"id":"2605.29089","kind":"arxiv","version":1}},"canonical_sha256":"a16bdf22549a79733c08f26c9ea00b0942b184fa0fccf87e610be8a04cf86865","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a16bdf22549a79733c08f26c9ea00b0942b184fa0fccf87e610be8a04cf86865","first_computed_at":"2026-05-29T01:05:17.696355Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-29T01:05:17.696355Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"mwKAFnz9jr3/gLzN9UgdKl9I8Q6hhSfdVgiRP97CiiyOl/GEFW3PWdFfj1ckTmT+diUv5bcXul2VyKRkjgY4CQ==","signature_status":"signed_v1","signed_at":"2026-05-29T01:05:17.697209Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.29089","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3fc6d2b8e8e5db66d70bd3faaed10b95ddfc36705b2e849dae759a27b7dab3bf","sha256:5f7300521988358717b052570af4e08c524d96c489144531fc3ac9895663e985"],"state_sha256":"dfc0dec01f6cb9f0bdb171ee274d88758d9c3f57f9f1ec5c8617d19dbc6e2f0c"}