{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:FM6LH3J2ZRHSLBDWYX4OW42DLJ","short_pith_number":"pith:FM6LH3J2","schema_version":"1.0","canonical_sha256":"2b3cb3ed3acc4f258476c5f8eb73435a75e034c5c42ef40f88c618ff81594c6e","source":{"kind":"arxiv","id":"2302.03686","version":2},"attestation_state":"computed","paper":{"title":"Long Horizon Temperature Scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andy Shih, Dorsa Sadigh, Stefano Ermon","submitted_at":"2023-02-07T18:59:32Z","abstract_excerpt":"Temperature scaling is a popular technique for tuning the sharpness of a model distribution. It is used extensively for sampling likely generations and calibrating model uncertainty, and even features as a controllable parameter to many large language models in deployment. However, autoregressive models rely on myopic temperature scaling that greedily optimizes the next token. To address this, we propose Long Horizon Temperature Scaling (LHTS), a novel approach for sampling from temperature-scaled joint distributions. LHTS is compatible with all likelihood-based models, and optimizes for the l"},"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":"2302.03686","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-02-07T18:59:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"5da329ee9e34fa0d65e29a99962e0b0f2be3c24d24f244f98a5b5af62f716cac","abstract_canon_sha256":"ce6a5cc6edaff96a56c13daf008f025d9d24a7778a6f22222c25b6138c835f48"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:55:51.938913Z","signature_b64":"1/SoEV1/YyWqRspSQhAdZKTtGMrD5RiqOHHodGnxN9esM5Qk4SgCTmgx+CWx0vJfwo+AzQ9JUQtFBkjiePP/Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2b3cb3ed3acc4f258476c5f8eb73435a75e034c5c42ef40f88c618ff81594c6e","last_reissued_at":"2026-07-05T06:55:51.938523Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:55:51.938523Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Long Horizon Temperature Scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andy Shih, Dorsa Sadigh, Stefano Ermon","submitted_at":"2023-02-07T18:59:32Z","abstract_excerpt":"Temperature scaling is a popular technique for tuning the sharpness of a model distribution. It is used extensively for sampling likely generations and calibrating model uncertainty, and even features as a controllable parameter to many large language models in deployment. However, autoregressive models rely on myopic temperature scaling that greedily optimizes the next token. To address this, we propose Long Horizon Temperature Scaling (LHTS), a novel approach for sampling from temperature-scaled joint distributions. LHTS is compatible with all likelihood-based models, and optimizes for the l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.03686","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/2302.03686/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":"2302.03686","created_at":"2026-07-05T06:55:51.938578+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.03686v2","created_at":"2026-07-05T06:55:51.938578+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.03686","created_at":"2026-07-05T06:55:51.938578+00:00"},{"alias_kind":"pith_short_12","alias_value":"FM6LH3J2ZRHS","created_at":"2026-07-05T06:55:51.938578+00:00"},{"alias_kind":"pith_short_16","alias_value":"FM6LH3J2ZRHSLBDW","created_at":"2026-07-05T06:55:51.938578+00:00"},{"alias_kind":"pith_short_8","alias_value":"FM6LH3J2","created_at":"2026-07-05T06:55:51.938578+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/FM6LH3J2ZRHSLBDWYX4OW42DLJ","json":"https://pith.science/pith/FM6LH3J2ZRHSLBDWYX4OW42DLJ.json","graph_json":"https://pith.science/api/pith-number/FM6LH3J2ZRHSLBDWYX4OW42DLJ/graph.json","events_json":"https://pith.science/api/pith-number/FM6LH3J2ZRHSLBDWYX4OW42DLJ/events.json","paper":"https://pith.science/paper/FM6LH3J2"},"agent_actions":{"view_html":"https://pith.science/pith/FM6LH3J2ZRHSLBDWYX4OW42DLJ","download_json":"https://pith.science/pith/FM6LH3J2ZRHSLBDWYX4OW42DLJ.json","view_paper":"https://pith.science/paper/FM6LH3J2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.03686&json=true","fetch_graph":"https://pith.science/api/pith-number/FM6LH3J2ZRHSLBDWYX4OW42DLJ/graph.json","fetch_events":"https://pith.science/api/pith-number/FM6LH3J2ZRHSLBDWYX4OW42DLJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FM6LH3J2ZRHSLBDWYX4OW42DLJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FM6LH3J2ZRHSLBDWYX4OW42DLJ/action/storage_attestation","attest_author":"https://pith.science/pith/FM6LH3J2ZRHSLBDWYX4OW42DLJ/action/author_attestation","sign_citation":"https://pith.science/pith/FM6LH3J2ZRHSLBDWYX4OW42DLJ/action/citation_signature","submit_replication":"https://pith.science/pith/FM6LH3J2ZRHSLBDWYX4OW42DLJ/action/replication_record"}},"created_at":"2026-07-05T06:55:51.938578+00:00","updated_at":"2026-07-05T06:55:51.938578+00:00"}