{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:X4275CXQSYDHADXRXDP6VY4L77","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":"6b06e441d66c1e1b848b1eb51290d264627b90c3a92d2896d7a819ac940670c7","cross_cats_sorted":["cs.IT","math.IT","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-02T08:40:07Z","title_canon_sha256":"8dc3f2d90b66ffdf710896d306b59e3687f09670d1358953aa40b9ffa61cf176"},"schema_version":"1.0","source":{"id":"2404.01730","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.01730","created_at":"2026-07-05T08:03:18Z"},{"alias_kind":"arxiv_version","alias_value":"2404.01730v1","created_at":"2026-07-05T08:03:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.01730","created_at":"2026-07-05T08:03:18Z"},{"alias_kind":"pith_short_12","alias_value":"X4275CXQSYDH","created_at":"2026-07-05T08:03:18Z"},{"alias_kind":"pith_short_16","alias_value":"X4275CXQSYDHADXR","created_at":"2026-07-05T08:03:18Z"},{"alias_kind":"pith_short_8","alias_value":"X4275CXQ","created_at":"2026-07-05T08:03:18Z"}],"graph_snapshots":[{"event_id":"sha256:1c055eded43918d600268574499b4e905bf47c3a6f26053554fe024d3426d53b","target":"graph","created_at":"2026-07-05T08:03:18Z","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/2404.01730/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Let $p$ denote a generative language model. Let $r$ denote a reward model that returns a scalar that captures the degree at which a draw from $p$ is preferred. The goal of language model alignment is to alter $p$ to a new distribution $\\phi$ that results in a higher expected reward while keeping $\\phi$ close to $p.$ A popular alignment method is the KL-constrained reinforcement learning (RL), which chooses a distribution $\\phi_\\Delta$ that maximizes $E_{\\phi_{\\Delta}} r(y)$ subject to a relative entropy constraint $KL(\\phi_\\Delta || p) \\leq \\Delta.$ Another simple alignment method is best-of-$","authors_text":"Ahmad Beirami, Ananda Theertha Suresh, Joy Qiping Yang, Salman Salamatian, Ziteng Sun","cross_cats":["cs.IT","math.IT","stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-02T08:40:07Z","title":"Asymptotics of Language Model Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.01730","kind":"arxiv","version":1},"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":"sha256:42c6e10b9f6519065543387199355d8a2925b2a0f4df63fdf474dfbfb03d845b","target":"record","created_at":"2026-07-05T08:03:18Z","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":"6b06e441d66c1e1b848b1eb51290d264627b90c3a92d2896d7a819ac940670c7","cross_cats_sorted":["cs.IT","math.IT","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-02T08:40:07Z","title_canon_sha256":"8dc3f2d90b66ffdf710896d306b59e3687f09670d1358953aa40b9ffa61cf176"},"schema_version":"1.0","source":{"id":"2404.01730","kind":"arxiv","version":1}},"canonical_sha256":"bf35fe8af09606700ef1b8dfeae38bffee4f6bbabdc315aa4aaacebc440ed4c9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"bf35fe8af09606700ef1b8dfeae38bffee4f6bbabdc315aa4aaacebc440ed4c9","first_computed_at":"2026-07-05T08:03:18.891229Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:03:18.891229Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"XpZo3bdsE0EV1bZvHkLlVGwNKPeYOfLDq/hqYM1gu+QR64QMwn23rpYTPERllIVeQSl+RUTsX378eobHn84gDA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:03:18.891637Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.01730","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:42c6e10b9f6519065543387199355d8a2925b2a0f4df63fdf474dfbfb03d845b","sha256:1c055eded43918d600268574499b4e905bf47c3a6f26053554fe024d3426d53b"],"state_sha256":"c24ea6296b25b7efce5abf5819b99ac00a34efcd7db4233e9823cc47a96a3d42"}