{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:GRI5B4Q77J7RAYHLX6BFMH4LEU","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":"68a7625169e50d32484d64941048ac686030b2f5241ffc6812c7769798fb64f1","cross_cats_sorted":["cs.CL","cs.IT","math.IT"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-03T18:39:13Z","title_canon_sha256":"ae3d87fcff3fce86fe653e1946f577b15e47e1cef781f81523abfaaf3a514983"},"schema_version":"1.0","source":{"id":"2401.01879","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.01879","created_at":"2026-07-05T11:11:30Z"},{"alias_kind":"arxiv_version","alias_value":"2401.01879v3","created_at":"2026-07-05T11:11:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.01879","created_at":"2026-07-05T11:11:30Z"},{"alias_kind":"pith_short_12","alias_value":"GRI5B4Q77J7R","created_at":"2026-07-05T11:11:30Z"},{"alias_kind":"pith_short_16","alias_value":"GRI5B4Q77J7RAYHL","created_at":"2026-07-05T11:11:30Z"},{"alias_kind":"pith_short_8","alias_value":"GRI5B4Q7","created_at":"2026-07-05T11:11:30Z"}],"graph_snapshots":[{"event_id":"sha256:72faeee95a011d777e38fc66fe4dc431b5905540fa89935d8c8dd3965d907aeb","target":"graph","created_at":"2026-07-05T11:11:30Z","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/2401.01879/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"A simple and effective method for the inference-time alignment and scaling test-time compute of generative models is best-of-$n$ sampling, where $n$ samples are drawn from a reference policy, ranked based on a reward function, and the highest ranking one is selected. A commonly used analytical expression in the literature claims that the KL divergence between the best-of-$n$ policy and the reference policy is equal to $\\log (n) - (n-1)/n.$ We disprove the validity of this claim, and show that it is an upper bound on the actual KL divergence. We also explore the tightness of this upper bound in","authors_text":"Ahmad Beirami, Alekh Agarwal, Alexander D'Amour, Ananda Theertha Suresh, Chirag Nagpal, Jacob Eisenstein, Jonathan Berant","cross_cats":["cs.CL","cs.IT","math.IT"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-03T18:39:13Z","title":"Theoretical guarantees on the best-of-n alignment policy"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.01879","kind":"arxiv","version":3},"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:a991b329c6e44218154a5305bfa6cc1f4b1fd0ebc3d8129af5760a81e1daee09","target":"record","created_at":"2026-07-05T11:11:30Z","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":"68a7625169e50d32484d64941048ac686030b2f5241ffc6812c7769798fb64f1","cross_cats_sorted":["cs.CL","cs.IT","math.IT"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-03T18:39:13Z","title_canon_sha256":"ae3d87fcff3fce86fe653e1946f577b15e47e1cef781f81523abfaaf3a514983"},"schema_version":"1.0","source":{"id":"2401.01879","kind":"arxiv","version":3}},"canonical_sha256":"3451d0f21ffa7f1060ebbf82561f8b250d934a1ee02c91541c122799a3d6eee9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3451d0f21ffa7f1060ebbf82561f8b250d934a1ee02c91541c122799a3d6eee9","first_computed_at":"2026-07-05T11:11:30.709019Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:11:30.709019Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"6uUg6PLC5hsOOcNh1wnfrvT9F+ywMbS72rym0PSUq9hjlzLLDTjsTfpo00OlXbcI+fyS8pU2HcflgH7WSGfgAA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:11:30.709554Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.01879","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a991b329c6e44218154a5305bfa6cc1f4b1fd0ebc3d8129af5760a81e1daee09","sha256:72faeee95a011d777e38fc66fe4dc431b5905540fa89935d8c8dd3965d907aeb"],"state_sha256":"52f91cb5c4af147b9034a4d9a4130d30d1f42eb39c1876d0b710b56fdeaa711c"}