{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:CZ2YV4EM5277FB6RZJZXUO4LFP","short_pith_number":"pith:CZ2YV4EM","schema_version":"1.0","canonical_sha256":"16758af08ceebff287d1ca737a3b8b2be68434e5cabcb78cb26cac193bb9105c","source":{"kind":"arxiv","id":"2607.27888","version":1},"attestation_state":"computed","paper":{"title":"Not All Tokens Deserve Equal Credit: Counterfactual Sensitivity Credit Reallocation for Long-CoT Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Qiangqiang He, Zhongheng Wu, Zijian Wang","submitted_at":"2026-07-30T09:03:33Z","abstract_excerpt":"Reinforcement learning with verifiable rewards (RLVR) is central to improving long-CoT reasoning in large language models. Critic-free methods such as GRPO convert response-level rewards into advantages and uniformly broadcast them across tokens, overlooking their unequal contributions to the final outcome. On-policy self-distillation (OPSD) instead provides dense distributional supervision by minimizing the forward KL divergence between an unprivileged policy and a privileged self-teacher, implicitly assuming that the resulting likelihood shifts encode reliable answer-aligned information. We "},"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":"2607.27888","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-07-30T09:03:33Z","cross_cats_sorted":[],"title_canon_sha256":"605c0a3badf5c256b6c1f8481a7aa2bf07ced061fef454e684b6822d843ae9b8","abstract_canon_sha256":"c936fe9a143a2130a48151f1935d0263d70983efd66b3dfa0d67d49c987e8b6c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"16758af08ceebff287d1ca737a3b8b2be68434e5cabcb78cb26cac193bb9105c","last_reissued_at":"2026-07-31T01:34:34.623236Z","signature_status":"unsigned_v0","first_computed_at":"2026-07-31T01:34:34.623236Z"},"graph_snapshot":{"paper":{"title":"Not All Tokens Deserve Equal Credit: Counterfactual Sensitivity Credit Reallocation for Long-CoT Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Qiangqiang He, Zhongheng Wu, Zijian Wang","submitted_at":"2026-07-30T09:03:33Z","abstract_excerpt":"Reinforcement learning with verifiable rewards (RLVR) is central to improving long-CoT reasoning in large language models. Critic-free methods such as GRPO convert response-level rewards into advantages and uniformly broadcast them across tokens, overlooking their unequal contributions to the final outcome. On-policy self-distillation (OPSD) instead provides dense distributional supervision by minimizing the forward KL divergence between an unprivileged policy and a privileged self-teacher, implicitly assuming that the resulting likelihood shifts encode reliable answer-aligned information. We "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.27888","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/2607.27888/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":"2607.27888","created_at":"2026-07-31T01:34:34.626385+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.27888v1","created_at":"2026-07-31T01:34:34.626385+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.27888","created_at":"2026-07-31T01:34:34.626385+00:00"},{"alias_kind":"pith_short_12","alias_value":"CZ2YV4EM5277","created_at":"2026-07-31T01:34:34.626385+00:00"},{"alias_kind":"pith_short_16","alias_value":"CZ2YV4EM5277FB6R","created_at":"2026-07-31T01:34:34.626385+00:00"},{"alias_kind":"pith_short_8","alias_value":"CZ2YV4EM","created_at":"2026-07-31T01:34:34.626385+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/CZ2YV4EM5277FB6RZJZXUO4LFP","json":"https://pith.science/pith/CZ2YV4EM5277FB6RZJZXUO4LFP.json","graph_json":"https://pith.science/api/pith-number/CZ2YV4EM5277FB6RZJZXUO4LFP/graph.json","events_json":"https://pith.science/api/pith-number/CZ2YV4EM5277FB6RZJZXUO4LFP/events.json","paper":"https://pith.science/paper/CZ2YV4EM"},"agent_actions":{"view_html":"https://pith.science/pith/CZ2YV4EM5277FB6RZJZXUO4LFP","download_json":"https://pith.science/pith/CZ2YV4EM5277FB6RZJZXUO4LFP.json","view_paper":"https://pith.science/paper/CZ2YV4EM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.27888&json=true","fetch_graph":"https://pith.science/api/pith-number/CZ2YV4EM5277FB6RZJZXUO4LFP/graph.json","fetch_events":"https://pith.science/api/pith-number/CZ2YV4EM5277FB6RZJZXUO4LFP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CZ2YV4EM5277FB6RZJZXUO4LFP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CZ2YV4EM5277FB6RZJZXUO4LFP/action/storage_attestation","attest_author":"https://pith.science/pith/CZ2YV4EM5277FB6RZJZXUO4LFP/action/author_attestation","sign_citation":"https://pith.science/pith/CZ2YV4EM5277FB6RZJZXUO4LFP/action/citation_signature","submit_replication":"https://pith.science/pith/CZ2YV4EM5277FB6RZJZXUO4LFP/action/replication_record"}},"created_at":"2026-07-31T01:34:34.626385+00:00","updated_at":"2026-07-31T01:34:34.626385+00:00"}