{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:32BWYDY7YPKDRBBBWEMVZQYZLJ","short_pith_number":"pith:32BWYDY7","canonical_record":{"source":{"id":"2305.14718","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-24T04:42:17Z","cross_cats_sorted":[],"title_canon_sha256":"39e8c84031ee33abedf83556362559bc466ffbaa41c8b588f0d8a50876044fb4","abstract_canon_sha256":"7d9eab610c6add09fcbd4e5df7bb688b40628ca3c686adccd07ac572e3d4cf19"},"schema_version":"1.0"},"canonical_sha256":"de836c0f1fc3d4388421b1195cc3195a7fd80b551fd190b9e4f15e50ec2bdd49","source":{"kind":"arxiv","id":"2305.14718","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.14718","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"arxiv_version","alias_value":"2305.14718v5","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.14718","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"pith_short_12","alias_value":"32BWYDY7YPKD","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"pith_short_16","alias_value":"32BWYDY7YPKDRBBB","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"pith_short_8","alias_value":"32BWYDY7","created_at":"2026-07-05T08:10:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:32BWYDY7YPKDRBBBWEMVZQYZLJ","target":"record","payload":{"canonical_record":{"source":{"id":"2305.14718","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-24T04:42:17Z","cross_cats_sorted":[],"title_canon_sha256":"39e8c84031ee33abedf83556362559bc466ffbaa41c8b588f0d8a50876044fb4","abstract_canon_sha256":"7d9eab610c6add09fcbd4e5df7bb688b40628ca3c686adccd07ac572e3d4cf19"},"schema_version":"1.0"},"canonical_sha256":"de836c0f1fc3d4388421b1195cc3195a7fd80b551fd190b9e4f15e50ec2bdd49","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:10:14.178002Z","signature_b64":"yvzxecu8OizhWvebkXqK29OjusOu6zwCUhbFt3SCit0s5Y4Gw8/vERyY4aVlP9+pYQAMM/HdGZED5wFI0KwJBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"de836c0f1fc3d4388421b1195cc3195a7fd80b551fd190b9e4f15e50ec2bdd49","last_reissued_at":"2026-07-05T08:10:14.177450Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:10:14.177450Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2305.14718","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:10:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mYeOKtAlrDJeAcozQHPlaSOqhcRdWqfyunYVB8uYBanfiWBAAFp1L2OvI/IANqHtJ4UjUj13jfv0tZ3+QHE1CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-16T23:09:13.259811Z"},"content_sha256":"af31a4599d71763660659e74e97297233f6d2aa433537e41e2295f64fda55511","schema_version":"1.0","event_id":"sha256:af31a4599d71763660659e74e97297233f6d2aa433537e41e2295f64fda55511"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:32BWYDY7YPKDRBBBWEMVZQYZLJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Leftover Lunch: Advantage-based Offline Reinforcement Learning for Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ashutosh Baheti, Faeze Brahman, Maarten Sap, Mark Riedl, Ronan Le Bras, Ximing Lu","submitted_at":"2023-05-24T04:42:17Z","abstract_excerpt":"Reinforcement Learning with Human Feedback (RLHF) is the most prominent method for Language Model (LM) alignment. However, RLHF is an unstable and data-hungry process that continually requires new high-quality LM-generated data for finetuning. We introduce Advantage-Leftover Lunch RL (A-LoL), a new class of offline policy gradient algorithms that enable RL training on any pre-existing data. By assuming the entire LM output sequence as a single action, A-LoL allows incorporating sequence-level classifiers or human-designed scoring functions as rewards. Subsequently, by using LM's value estimate"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.14718","kind":"arxiv","version":5},"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/2305.14718/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:10:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fN+a1HR8VW9qpPS40/9p4rdkm/23a0+74ofQLF0I+Z17Xu3g/GZDcc3xwIEtR/NSv2sXAzYD8FN09vWumRT1Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-16T23:09:13.260308Z"},"content_sha256":"89cadd436c6c8f52f940c789644e63687228ee47958d315528827898747a3093","schema_version":"1.0","event_id":"sha256:89cadd436c6c8f52f940c789644e63687228ee47958d315528827898747a3093"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/32BWYDY7YPKDRBBBWEMVZQYZLJ/bundle.json","state_url":"https://pith.science/pith/32BWYDY7YPKDRBBBWEMVZQYZLJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/32BWYDY7YPKDRBBBWEMVZQYZLJ/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-16T23:09:13Z","links":{"resolver":"https://pith.science/pith/32BWYDY7YPKDRBBBWEMVZQYZLJ","bundle":"https://pith.science/pith/32BWYDY7YPKDRBBBWEMVZQYZLJ/bundle.json","state":"https://pith.science/pith/32BWYDY7YPKDRBBBWEMVZQYZLJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/32BWYDY7YPKDRBBBWEMVZQYZLJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:32BWYDY7YPKDRBBBWEMVZQYZLJ","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":"7d9eab610c6add09fcbd4e5df7bb688b40628ca3c686adccd07ac572e3d4cf19","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-24T04:42:17Z","title_canon_sha256":"39e8c84031ee33abedf83556362559bc466ffbaa41c8b588f0d8a50876044fb4"},"schema_version":"1.0","source":{"id":"2305.14718","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2305.14718","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"arxiv_version","alias_value":"2305.14718v5","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.14718","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"pith_short_12","alias_value":"32BWYDY7YPKD","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"pith_short_16","alias_value":"32BWYDY7YPKDRBBB","created_at":"2026-07-05T08:10:14Z"},{"alias_kind":"pith_short_8","alias_value":"32BWYDY7","created_at":"2026-07-05T08:10:14Z"}],"graph_snapshots":[{"event_id":"sha256:89cadd436c6c8f52f940c789644e63687228ee47958d315528827898747a3093","target":"graph","created_at":"2026-07-05T08:10:14Z","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/2305.14718/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement Learning with Human Feedback (RLHF) is the most prominent method for Language Model (LM) alignment. However, RLHF is an unstable and data-hungry process that continually requires new high-quality LM-generated data for finetuning. We introduce Advantage-Leftover Lunch RL (A-LoL), a new class of offline policy gradient algorithms that enable RL training on any pre-existing data. By assuming the entire LM output sequence as a single action, A-LoL allows incorporating sequence-level classifiers or human-designed scoring functions as rewards. Subsequently, by using LM's value estimate","authors_text":"Ashutosh Baheti, Faeze Brahman, Maarten Sap, Mark Riedl, Ronan Le Bras, Ximing Lu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-24T04:42:17Z","title":"Leftover Lunch: Advantage-based Offline Reinforcement Learning for Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.14718","kind":"arxiv","version":5},"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:af31a4599d71763660659e74e97297233f6d2aa433537e41e2295f64fda55511","target":"record","created_at":"2026-07-05T08:10:14Z","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":"7d9eab610c6add09fcbd4e5df7bb688b40628ca3c686adccd07ac572e3d4cf19","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-24T04:42:17Z","title_canon_sha256":"39e8c84031ee33abedf83556362559bc466ffbaa41c8b588f0d8a50876044fb4"},"schema_version":"1.0","source":{"id":"2305.14718","kind":"arxiv","version":5}},"canonical_sha256":"de836c0f1fc3d4388421b1195cc3195a7fd80b551fd190b9e4f15e50ec2bdd49","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"de836c0f1fc3d4388421b1195cc3195a7fd80b551fd190b9e4f15e50ec2bdd49","first_computed_at":"2026-07-05T08:10:14.177450Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:10:14.177450Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"yvzxecu8OizhWvebkXqK29OjusOu6zwCUhbFt3SCit0s5Y4Gw8/vERyY4aVlP9+pYQAMM/HdGZED5wFI0KwJBg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:10:14.178002Z","signed_message":"canonical_sha256_bytes"},"source_id":"2305.14718","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:af31a4599d71763660659e74e97297233f6d2aa433537e41e2295f64fda55511","sha256:89cadd436c6c8f52f940c789644e63687228ee47958d315528827898747a3093"],"state_sha256":"13c4553a83c45d4fc790c8af8f6d0cbf013e7533435e1475f7b8dd932e9bbfb9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"H1gsjzANUPuuMhuBqeKL9gJDFyp6eQcArHMbklG8K7NLokUWYnLPUwEyqs5ywTaKbSj5VMLqxdtfS7ocd4wfDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-16T23:09:13.266525Z","bundle_sha256":"c89be848fc9b2f79cf0b421cf4e3080e62337c88d2b998269e5cdb6656281eef"}}