{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:HLJ66AEPRSWHDQPNCYOLVAFXA3","short_pith_number":"pith:HLJ66AEP","schema_version":"1.0","canonical_sha256":"3ad3ef008f8cac71c1ed161cba80b706ca5afea0399c8163a820ef63656dadbf","source":{"kind":"arxiv","id":"2602.19049","version":2},"attestation_state":"computed","paper":{"title":"IAPO: Information-Aware Policy Optimization for Token-Efficient Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Jundong Li, Lin Su, Mingjia Shi, Qi Guo, Wendy Zheng, Xiaoqing Wang, Yaochen Zhu, Yinhan He","submitted_at":"2026-02-22T05:30:14Z","abstract_excerpt":"Large language models increasingly rely on long chains of thought to improve accuracy, yet such gains come with substantial inference-time costs. We revisit token-efficient post-training and argue that existing sequence-level reward-shaping methods offer limited control over how reasoning effort is allocated across tokens. To bridge the gap, we propose IAPO, an information-theoretic post-training framework that assigns token-wise advantages based on each token's conditional mutual information (MI) with the final answer. This yields an explicit, principled mechanism for identifying informative "},"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":"2602.19049","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-02-22T05:30:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"14d1161a84a31e11fc4717168c8706b45dc2ed983f33bce314ed9462580c9300","abstract_canon_sha256":"1101fe7620a5a170ae019106d4ea63b3229fa982abd9f3482e3c7b62ee3c3482"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T01:02:35.085863Z","signature_b64":"I6MbZRWcHf/B8pfplWIXZREv7sZbuO/FLP63/9iEuYlT7dUCrRvH7M7ep8zBk8uDnWew3ql9twfjRVGnN3kWDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3ad3ef008f8cac71c1ed161cba80b706ca5afea0399c8163a820ef63656dadbf","last_reissued_at":"2026-06-01T01:02:35.084547Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T01:02:35.084547Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"IAPO: Information-Aware Policy Optimization for Token-Efficient Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Jundong Li, Lin Su, Mingjia Shi, Qi Guo, Wendy Zheng, Xiaoqing Wang, Yaochen Zhu, Yinhan He","submitted_at":"2026-02-22T05:30:14Z","abstract_excerpt":"Large language models increasingly rely on long chains of thought to improve accuracy, yet such gains come with substantial inference-time costs. We revisit token-efficient post-training and argue that existing sequence-level reward-shaping methods offer limited control over how reasoning effort is allocated across tokens. To bridge the gap, we propose IAPO, an information-theoretic post-training framework that assigns token-wise advantages based on each token's conditional mutual information (MI) with the final answer. This yields an explicit, principled mechanism for identifying informative "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.19049","kind":"arxiv","version":2},"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/2602.19049/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":"2602.19049","created_at":"2026-06-01T01:02:35.084746+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.19049v2","created_at":"2026-06-01T01:02:35.084746+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.19049","created_at":"2026-06-01T01:02:35.084746+00:00"},{"alias_kind":"pith_short_12","alias_value":"HLJ66AEPRSWH","created_at":"2026-06-01T01:02:35.084746+00:00"},{"alias_kind":"pith_short_16","alias_value":"HLJ66AEPRSWHDQPN","created_at":"2026-06-01T01:02:35.084746+00:00"},{"alias_kind":"pith_short_8","alias_value":"HLJ66AEP","created_at":"2026-06-01T01:02:35.084746+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/HLJ66AEPRSWHDQPNCYOLVAFXA3","json":"https://pith.science/pith/HLJ66AEPRSWHDQPNCYOLVAFXA3.json","graph_json":"https://pith.science/api/pith-number/HLJ66AEPRSWHDQPNCYOLVAFXA3/graph.json","events_json":"https://pith.science/api/pith-number/HLJ66AEPRSWHDQPNCYOLVAFXA3/events.json","paper":"https://pith.science/paper/HLJ66AEP"},"agent_actions":{"view_html":"https://pith.science/pith/HLJ66AEPRSWHDQPNCYOLVAFXA3","download_json":"https://pith.science/pith/HLJ66AEPRSWHDQPNCYOLVAFXA3.json","view_paper":"https://pith.science/paper/HLJ66AEP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.19049&json=true","fetch_graph":"https://pith.science/api/pith-number/HLJ66AEPRSWHDQPNCYOLVAFXA3/graph.json","fetch_events":"https://pith.science/api/pith-number/HLJ66AEPRSWHDQPNCYOLVAFXA3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HLJ66AEPRSWHDQPNCYOLVAFXA3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HLJ66AEPRSWHDQPNCYOLVAFXA3/action/storage_attestation","attest_author":"https://pith.science/pith/HLJ66AEPRSWHDQPNCYOLVAFXA3/action/author_attestation","sign_citation":"https://pith.science/pith/HLJ66AEPRSWHDQPNCYOLVAFXA3/action/citation_signature","submit_replication":"https://pith.science/pith/HLJ66AEPRSWHDQPNCYOLVAFXA3/action/replication_record"}},"created_at":"2026-06-01T01:02:35.084746+00:00","updated_at":"2026-06-01T01:02:35.084746+00:00"}