{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QHURDU5EXR3Z33UPGC5KGR4WD3","short_pith_number":"pith:QHURDU5E","schema_version":"1.0","canonical_sha256":"81e911d3a4bc779dee8f30baa347961ed891e79c082fb91c4095c1ee058690a2","source":{"kind":"arxiv","id":"2605.30844","version":1},"attestation_state":"computed","paper":{"title":"Fine-Tuning Improves Information Conveyance in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.CL","authors_text":"Haifeng Xu, Weiyi Tian, Yuwei Cheng","submitted_at":"2026-05-29T05:05:01Z","abstract_excerpt":"Fine-tuning is often believed to reduce uncertainty and diversity in large language models, but existing analyses overlook output length, a key confounder, and therefore fail to capture how uncertainty is distributed across an entire generation rollout. To address this, we propose Canopy Entropy ($\\mathrm{CE}^\\star$), a measure that views language generation from a tree perspective, where ``canopy'' represents the space of all possible rollouts, making $\\mathrm{CE}^\\star$ naturally quantify the effective size of the generation space. $\\mathrm{CE}^\\star$ jointly captures uncertainty in both the"},"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":"2605.30844","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-29T05:05:01Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"80a5394460b761c70465ac3bd681ae733dc3eff152b04736a896d290d488ee1f","abstract_canon_sha256":"20a6cdb36a39340cc1a04321be3df98dccbcbc71842f8807925005808e031ab7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T01:03:20.596246Z","signature_b64":"UuzcBuoGbArZlRTtoMruYSxO6gS9Vj9N6mwQuuF0BjdAloPHDI7C+4ljkHtiTYdC3xha1Oq6OmGH/5i98yyDCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"81e911d3a4bc779dee8f30baa347961ed891e79c082fb91c4095c1ee058690a2","last_reissued_at":"2026-06-01T01:03:20.595156Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T01:03:20.595156Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fine-Tuning Improves Information Conveyance in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.CL","authors_text":"Haifeng Xu, Weiyi Tian, Yuwei Cheng","submitted_at":"2026-05-29T05:05:01Z","abstract_excerpt":"Fine-tuning is often believed to reduce uncertainty and diversity in large language models, but existing analyses overlook output length, a key confounder, and therefore fail to capture how uncertainty is distributed across an entire generation rollout. To address this, we propose Canopy Entropy ($\\mathrm{CE}^\\star$), a measure that views language generation from a tree perspective, where ``canopy'' represents the space of all possible rollouts, making $\\mathrm{CE}^\\star$ naturally quantify the effective size of the generation space. $\\mathrm{CE}^\\star$ jointly captures uncertainty in both the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30844","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/2605.30844/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":"2605.30844","created_at":"2026-06-01T01:03:20.595306+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.30844v1","created_at":"2026-06-01T01:03:20.595306+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30844","created_at":"2026-06-01T01:03:20.595306+00:00"},{"alias_kind":"pith_short_12","alias_value":"QHURDU5EXR3Z","created_at":"2026-06-01T01:03:20.595306+00:00"},{"alias_kind":"pith_short_16","alias_value":"QHURDU5EXR3Z33UP","created_at":"2026-06-01T01:03:20.595306+00:00"},{"alias_kind":"pith_short_8","alias_value":"QHURDU5E","created_at":"2026-06-01T01:03:20.595306+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/QHURDU5EXR3Z33UPGC5KGR4WD3","json":"https://pith.science/pith/QHURDU5EXR3Z33UPGC5KGR4WD3.json","graph_json":"https://pith.science/api/pith-number/QHURDU5EXR3Z33UPGC5KGR4WD3/graph.json","events_json":"https://pith.science/api/pith-number/QHURDU5EXR3Z33UPGC5KGR4WD3/events.json","paper":"https://pith.science/paper/QHURDU5E"},"agent_actions":{"view_html":"https://pith.science/pith/QHURDU5EXR3Z33UPGC5KGR4WD3","download_json":"https://pith.science/pith/QHURDU5EXR3Z33UPGC5KGR4WD3.json","view_paper":"https://pith.science/paper/QHURDU5E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.30844&json=true","fetch_graph":"https://pith.science/api/pith-number/QHURDU5EXR3Z33UPGC5KGR4WD3/graph.json","fetch_events":"https://pith.science/api/pith-number/QHURDU5EXR3Z33UPGC5KGR4WD3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QHURDU5EXR3Z33UPGC5KGR4WD3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QHURDU5EXR3Z33UPGC5KGR4WD3/action/storage_attestation","attest_author":"https://pith.science/pith/QHURDU5EXR3Z33UPGC5KGR4WD3/action/author_attestation","sign_citation":"https://pith.science/pith/QHURDU5EXR3Z33UPGC5KGR4WD3/action/citation_signature","submit_replication":"https://pith.science/pith/QHURDU5EXR3Z33UPGC5KGR4WD3/action/replication_record"}},"created_at":"2026-06-01T01:03:20.595306+00:00","updated_at":"2026-06-01T01:03:20.595306+00:00"}