{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:PMVQPIL2OD3CIMYDZUJND2I4GO","short_pith_number":"pith:PMVQPIL2","canonical_record":{"source":{"id":"2204.14146","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-29T15:06:58Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"f636a12d18e92c8dc8c73d077f642da2d552c7b817902c9d0773b513dde36d41","abstract_canon_sha256":"35b8ac2c0782583e606db838a3786a6858ed893e51a220ba660ee5839d2335fd"},"schema_version":"1.0"},"canonical_sha256":"7b2b07a17a70f6243303cd12d1e91c339f272b50d6651a412fb24a8624a34329","source":{"kind":"arxiv","id":"2204.14146","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2204.14146","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"arxiv_version","alias_value":"2204.14146v4","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.14146","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"pith_short_12","alias_value":"PMVQPIL2OD3C","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"pith_short_16","alias_value":"PMVQPIL2OD3CIMYD","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"pith_short_8","alias_value":"PMVQPIL2","created_at":"2026-07-05T05:17:01Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:PMVQPIL2OD3CIMYDZUJND2I4GO","target":"record","payload":{"canonical_record":{"source":{"id":"2204.14146","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-29T15:06:58Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"f636a12d18e92c8dc8c73d077f642da2d552c7b817902c9d0773b513dde36d41","abstract_canon_sha256":"35b8ac2c0782583e606db838a3786a6858ed893e51a220ba660ee5839d2335fd"},"schema_version":"1.0"},"canonical_sha256":"7b2b07a17a70f6243303cd12d1e91c339f272b50d6651a412fb24a8624a34329","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:17:01.867828Z","signature_b64":"UCWbUl+fU3yX0EsMzjkq3S52mGxxj7gxjTz0+GoiNy5i/eE8fxKaTDnaBCNveTzIPsb+fwseUB4J2FD4OcZNAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b2b07a17a70f6243303cd12d1e91c339f272b50d6651a412fb24a8624a34329","last_reissued_at":"2026-07-05T05:17:01.867374Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:17:01.867374Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2204.14146","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T05:17:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9qglLC4JcVUE0XqqVH5Z4sDcdKkZjQtIwOQxFD/5wxzGgahFr5wa2RPCTZpHy/X/p1aojHHlKKAEgh75Y/ArAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T23:45:17.455930Z"},"content_sha256":"b54e4ebdcdab86ec5494b3995c2cf2a2247a67efa796b51bfa985df89468a96e","schema_version":"1.0","event_id":"sha256:b54e4ebdcdab86ec5494b3995c2cf2a2247a67efa796b51bfa985df89468a96e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:PMVQPIL2OD3CIMYDZUJND2I4GO","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Training Language Models with Language Feedback","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Angelica Chen, Ethan Perez, J\\'er\\'emy Scheurer, Jon Ander Campos, Jun Shern Chan, Kyunghyun Cho","submitted_at":"2022-04-29T15:06:58Z","abstract_excerpt":"Pretrained language models often do not perform tasks in ways that are in line with our preferences, e.g., generating offensive text or factually incorrect summaries. Recent work approaches the above issue by learning from a simple form of human evaluation: comparisons between pairs of model-generated task outputs. Comparison feedback conveys limited information about human preferences per human evaluation. Here, we propose to learn from natural language feedback, which conveys more information per human evaluation. We learn from language feedback on model outputs using a three-step learning a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.14146","kind":"arxiv","version":4},"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/2204.14146/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":"2026-07-05T05:17:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RiFmDwp7xZJIn2MmZ/Mff2+KAapm0Zez+wa5uvnRe3LLOsWOB0qRkXzZf9hBIZp3DBPAlHaq8YQ7vTzhtIvSAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T23:45:17.456433Z"},"content_sha256":"f8795c48375f25d6e09a0cccf46b9bd9bb0704df2f88169f333c42539010d903","schema_version":"1.0","event_id":"sha256:f8795c48375f25d6e09a0cccf46b9bd9bb0704df2f88169f333c42539010d903"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/PMVQPIL2OD3CIMYDZUJND2I4GO/bundle.json","state_url":"https://pith.science/pith/PMVQPIL2OD3CIMYDZUJND2I4GO/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/PMVQPIL2OD3CIMYDZUJND2I4GO/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":"2026-08-09T23:45:17Z","links":{"resolver":"https://pith.science/pith/PMVQPIL2OD3CIMYDZUJND2I4GO","bundle":"https://pith.science/pith/PMVQPIL2OD3CIMYDZUJND2I4GO/bundle.json","state":"https://pith.science/pith/PMVQPIL2OD3CIMYDZUJND2I4GO/state.json","well_known_bundle":"https://pith.science/.well-known/pith/PMVQPIL2OD3CIMYDZUJND2I4GO/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:PMVQPIL2OD3CIMYDZUJND2I4GO","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":"35b8ac2c0782583e606db838a3786a6858ed893e51a220ba660ee5839d2335fd","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-29T15:06:58Z","title_canon_sha256":"f636a12d18e92c8dc8c73d077f642da2d552c7b817902c9d0773b513dde36d41"},"schema_version":"1.0","source":{"id":"2204.14146","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2204.14146","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"arxiv_version","alias_value":"2204.14146v4","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.14146","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"pith_short_12","alias_value":"PMVQPIL2OD3C","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"pith_short_16","alias_value":"PMVQPIL2OD3CIMYD","created_at":"2026-07-05T05:17:01Z"},{"alias_kind":"pith_short_8","alias_value":"PMVQPIL2","created_at":"2026-07-05T05:17:01Z"}],"graph_snapshots":[{"event_id":"sha256:f8795c48375f25d6e09a0cccf46b9bd9bb0704df2f88169f333c42539010d903","target":"graph","created_at":"2026-07-05T05:17:01Z","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/2204.14146/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pretrained language models often do not perform tasks in ways that are in line with our preferences, e.g., generating offensive text or factually incorrect summaries. Recent work approaches the above issue by learning from a simple form of human evaluation: comparisons between pairs of model-generated task outputs. Comparison feedback conveys limited information about human preferences per human evaluation. Here, we propose to learn from natural language feedback, which conveys more information per human evaluation. We learn from language feedback on model outputs using a three-step learning a","authors_text":"Angelica Chen, Ethan Perez, J\\'er\\'emy Scheurer, Jon Ander Campos, Jun Shern Chan, Kyunghyun Cho","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-29T15:06:58Z","title":"Training Language Models with Language Feedback"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.14146","kind":"arxiv","version":4},"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:b54e4ebdcdab86ec5494b3995c2cf2a2247a67efa796b51bfa985df89468a96e","target":"record","created_at":"2026-07-05T05:17:01Z","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":"35b8ac2c0782583e606db838a3786a6858ed893e51a220ba660ee5839d2335fd","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-29T15:06:58Z","title_canon_sha256":"f636a12d18e92c8dc8c73d077f642da2d552c7b817902c9d0773b513dde36d41"},"schema_version":"1.0","source":{"id":"2204.14146","kind":"arxiv","version":4}},"canonical_sha256":"7b2b07a17a70f6243303cd12d1e91c339f272b50d6651a412fb24a8624a34329","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7b2b07a17a70f6243303cd12d1e91c339f272b50d6651a412fb24a8624a34329","first_computed_at":"2026-07-05T05:17:01.867374Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:17:01.867374Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"UCWbUl+fU3yX0EsMzjkq3S52mGxxj7gxjTz0+GoiNy5i/eE8fxKaTDnaBCNveTzIPsb+fwseUB4J2FD4OcZNAA==","signature_status":"signed_v1","signed_at":"2026-07-05T05:17:01.867828Z","signed_message":"canonical_sha256_bytes"},"source_id":"2204.14146","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b54e4ebdcdab86ec5494b3995c2cf2a2247a67efa796b51bfa985df89468a96e","sha256:f8795c48375f25d6e09a0cccf46b9bd9bb0704df2f88169f333c42539010d903"],"state_sha256":"bfd580b79883e1557738c9dc7d91418a241fe20c59e1b9f77e9ce5a8b2ea798f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oxJFsvMArsaTYUjCvnTwCbjxZEkGMG/h9rj07LpVOR6ip9v4b6vmL2OnBUoo4XElUkOD00rtmB5muSRtULc3CA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T23:45:17.460724Z","bundle_sha256":"c2a272ce5b5ef6e7a3ecf390e5f66abf5d4b42105429b06bc012db1c67b36c8d"}}