{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:W5AW73QHXHRS4ZAC566MZ6OMRR","short_pith_number":"pith:W5AW73QH","canonical_record":{"source":{"id":"2203.11147","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:26:29Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"13c59520008b99b844030d7803e45e1caa24dcb0fc66e41169eb41631e273ec5","abstract_canon_sha256":"3f27c4af1e562a931e61026b868ea299ed8a7336b9e159d0dbe738400d973823"},"schema_version":"1.0"},"canonical_sha256":"b7416fee07b9e32e6402efbcccf9cc8c75905c39fdbe93293cc1ef5a6e1101d5","source":{"kind":"arxiv","id":"2203.11147","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2203.11147","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"arxiv_version","alias_value":"2203.11147v1","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.11147","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"pith_short_12","alias_value":"W5AW73QHXHRS","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"W5AW73QHXHRS4ZAC","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"W5AW73QH","created_at":"2026-05-18T12:33:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:W5AW73QHXHRS4ZAC566MZ6OMRR","target":"record","payload":{"canonical_record":{"source":{"id":"2203.11147","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:26:29Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"13c59520008b99b844030d7803e45e1caa24dcb0fc66e41169eb41631e273ec5","abstract_canon_sha256":"3f27c4af1e562a931e61026b868ea299ed8a7336b9e159d0dbe738400d973823"},"schema_version":"1.0"},"canonical_sha256":"b7416fee07b9e32e6402efbcccf9cc8c75905c39fdbe93293cc1ef5a6e1101d5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:14.206967Z","signature_b64":"op1DX+t2vqDDhyd+MoJdrNCcCF3FyRpCPZg7gOtjTgLWJfZofjprS7EdTMsJjC0n6Mz58Vr4mSqN0DpOKl2tCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b7416fee07b9e32e6402efbcccf9cc8c75905c39fdbe93293cc1ef5a6e1101d5","last_reissued_at":"2026-05-17T23:38:14.206266Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:14.206266Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2203.11147","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6m+91+YKQFe41JUFHOf4S6MRyX9A7BaPyGq4nytBaBfYZuzrVRyA24pcz3UTT/HSl2WIMbcoidOnqncSFnEvBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T22:11:44.314793Z"},"content_sha256":"728841da86dc62209d64ce03e96a06240b7235c3e889e23851ab60ff245aaa56","schema_version":"1.0","event_id":"sha256:728841da86dc62209d64ce03e96a06240b7235c3e889e23851ab60ff245aaa56"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:W5AW73QHXHRS4ZAC566MZ6OMRR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Teaching language models to support answers with verified quotes","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A 280 billion parameter model can be trained to answer questions with specific cited evidence from documents and to abstain when uncertain.","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Francis Song, Geoffrey Irving, Jacob Menick, John Aslanides, Lucy Campbell-Gillingham, Maja Trebacz, Martin Chadwick, Mia Glaese, Nat McAleese, Susannah Young, Vladimir Mikulik","submitted_at":"2022-03-21T17:26:29Z","abstract_excerpt":"Recent large language models often answer factual questions correctly. But users can't trust any given claim a model makes without fact-checking, because language models can hallucinate convincing nonsense. In this work we use reinforcement learning from human preferences (RLHP) to train \"open-book\" QA models that generate answers whilst also citing specific evidence for their claims, which aids in the appraisal of correctness. Supporting evidence is drawn from multiple documents found via a search engine, or from a single user-provided document. Our 280 billion parameter model, GopherCite, is"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our 280 billion parameter model, GopherCite, is able to produce answers with high quality supporting evidence and abstain from answering when unsure. The model's response is found to be high-quality 80% of the time on this Natural Questions subset, and 67% of the time on the ELI5 subset. Abstaining from the third of questions for which it is most unsure improves performance to 90% and 80% respectively.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That human raters' preferences for 'high quality supporting evidence' during RLHP training generalize to produce reliable citations and that the model's internal uncertainty signal for abstention is well-calibrated without introducing new biases.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"GopherCite produces answers with supporting evidence citations, rated high-quality 80% of the time on Natural Questions and 67% on ELI5, improving to 90% and 80% with abstention on uncertain questions.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A 280 billion parameter model can be trained to answer questions with specific cited evidence from documents and to abstain when uncertain.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"0bfbea845d9b4ea5c26a8c4ad7ed798d6d75e5d37e0a6c3c78e702e9d62020fd"},"source":{"id":"2203.11147","kind":"arxiv","version":1},"verdict":{"id":"9612394c-192c-4c1f-a9f9-d2f981f7f4f1","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T11:41:27.964902Z","strongest_claim":"Our 280 billion parameter model, GopherCite, is able to produce answers with high quality supporting evidence and abstain from answering when unsure. The model's response is found to be high-quality 80% of the time on this Natural Questions subset, and 67% of the time on the ELI5 subset. Abstaining from the third of questions for which it is most unsure improves performance to 90% and 80% respectively.","one_line_summary":"GopherCite produces answers with supporting evidence citations, rated high-quality 80% of the time on Natural Questions and 67% on ELI5, improving to 90% and 80% with abstention on uncertain questions.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That human raters' preferences for 'high quality supporting evidence' during RLHP training generalize to produce reliable citations and that the model's internal uncertainty signal for abstention is well-calibrated without introducing new biases.","pith_extraction_headline":"A 280 billion parameter model can be trained to answer questions with specific cited evidence from documents and to abstain when uncertain."},"references":{"count":14,"sample":[{"doi":"10.1145/3041021.3053375","year":2018,"title":"ISBN 9781450349147","work_id":"f9659ba7-d3b9-47d1-a957-64fed829351f","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1177/0894439317715434","year":2018,"title":"road draft tube","work_id":"be985a9d-c962-4bf8-91fa-284ca394a293","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"{url} • {claim} See this fragment from \"{title}\"[1]: {quote}","work_id":"5aad03f9-dc2e-4995-b1ce-97c186899a82","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"{quote}\" Source:","work_id":"23284550-9063-4b4d-90d9-825a613e8ccd","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"What happens if you smash a mirror?","work_id":"712292c7-2504-4989-b456-03e962093e6c","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":14,"snapshot_sha256":"3aa19ac14cbd371c1e693ce659656e65f10ba455726feb7f80c2706295bbbc85","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":"9612394c-192c-4c1f-a9f9-d2f981f7f4f1"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eOH3CZs6sJsujomHL+oK9X97ZL4Xf/+z3er2QrjAo9UnaJzjQsH3F6EWJ7dkg/H0afrbNi3hH5DhMlSItKTmDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T22:11:44.315943Z"},"content_sha256":"440e2e1282e943bb0d459eee5d7ae41d9c6ff75aff0bdf9e54f12cd0f1b5f316","schema_version":"1.0","event_id":"sha256:440e2e1282e943bb0d459eee5d7ae41d9c6ff75aff0bdf9e54f12cd0f1b5f316"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/W5AW73QHXHRS4ZAC566MZ6OMRR/bundle.json","state_url":"https://pith.science/pith/W5AW73QHXHRS4ZAC566MZ6OMRR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/W5AW73QHXHRS4ZAC566MZ6OMRR/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":"2026-05-26T22:11:44Z","links":{"resolver":"https://pith.science/pith/W5AW73QHXHRS4ZAC566MZ6OMRR","bundle":"https://pith.science/pith/W5AW73QHXHRS4ZAC566MZ6OMRR/bundle.json","state":"https://pith.science/pith/W5AW73QHXHRS4ZAC566MZ6OMRR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/W5AW73QHXHRS4ZAC566MZ6OMRR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:W5AW73QHXHRS4ZAC566MZ6OMRR","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":"3f27c4af1e562a931e61026b868ea299ed8a7336b9e159d0dbe738400d973823","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:26:29Z","title_canon_sha256":"13c59520008b99b844030d7803e45e1caa24dcb0fc66e41169eb41631e273ec5"},"schema_version":"1.0","source":{"id":"2203.11147","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2203.11147","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"arxiv_version","alias_value":"2203.11147v1","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.11147","created_at":"2026-05-17T23:38:14Z"},{"alias_kind":"pith_short_12","alias_value":"W5AW73QHXHRS","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"W5AW73QHXHRS4ZAC","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"W5AW73QH","created_at":"2026-05-18T12:33:33Z"}],"graph_snapshots":[{"event_id":"sha256:440e2e1282e943bb0d459eee5d7ae41d9c6ff75aff0bdf9e54f12cd0f1b5f316","target":"graph","created_at":"2026-05-17T23:38:14Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Our 280 billion parameter model, GopherCite, is able to produce answers with high quality supporting evidence and abstain from answering when unsure. The model's response is found to be high-quality 80% of the time on this Natural Questions subset, and 67% of the time on the ELI5 subset. Abstaining from the third of questions for which it is most unsure improves performance to 90% and 80% respectively."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That human raters' preferences for 'high quality supporting evidence' during RLHP training generalize to produce reliable citations and that the model's internal uncertainty signal for abstention is well-calibrated without introducing new biases."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"GopherCite produces answers with supporting evidence citations, rated high-quality 80% of the time on Natural Questions and 67% on ELI5, improving to 90% and 80% with abstention on uncertain questions."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A 280 billion parameter model can be trained to answer questions with specific cited evidence from documents and to abstain when uncertain."}],"snapshot_sha256":"0bfbea845d9b4ea5c26a8c4ad7ed798d6d75e5d37e0a6c3c78e702e9d62020fd"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Recent large language models often answer factual questions correctly. But users can't trust any given claim a model makes without fact-checking, because language models can hallucinate convincing nonsense. In this work we use reinforcement learning from human preferences (RLHP) to train \"open-book\" QA models that generate answers whilst also citing specific evidence for their claims, which aids in the appraisal of correctness. Supporting evidence is drawn from multiple documents found via a search engine, or from a single user-provided document. Our 280 billion parameter model, GopherCite, is","authors_text":"Francis Song, Geoffrey Irving, Jacob Menick, John Aslanides, Lucy Campbell-Gillingham, Maja Trebacz, Martin Chadwick, Mia Glaese, Nat McAleese, Susannah Young, Vladimir Mikulik","cross_cats":["cs.LG"],"headline":"A 280 billion parameter model can be trained to answer questions with specific cited evidence from documents and to abstain when uncertain.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:26:29Z","title":"Teaching language models to support answers with verified quotes"},"references":{"count":14,"internal_anchors":0,"resolved_work":14,"sample":[{"cited_arxiv_id":"","doi":"10.1145/3041021.3053375","is_internal_anchor":false,"ref_index":1,"title":"ISBN 9781450349147","work_id":"f9659ba7-d3b9-47d1-a957-64fed829351f","year":2018},{"cited_arxiv_id":"","doi":"10.1177/0894439317715434","is_internal_anchor":false,"ref_index":2,"title":"road draft tube","work_id":"be985a9d-c962-4bf8-91fa-284ca394a293","year":2018},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"{url} • {claim} See this fragment from \"{title}\"[1]: {quote}","work_id":"5aad03f9-dc2e-4995-b1ce-97c186899a82","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"{quote}\" Source:","work_id":"23284550-9063-4b4d-90d9-825a613e8ccd","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"What happens if you smash a mirror?","work_id":"712292c7-2504-4989-b456-03e962093e6c","year":2021}],"snapshot_sha256":"3aa19ac14cbd371c1e693ce659656e65f10ba455726feb7f80c2706295bbbc85"},"source":{"id":"2203.11147","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-17T11:41:27.964902Z","id":"9612394c-192c-4c1f-a9f9-d2f981f7f4f1","model_set":{"reader":"grok-4.3"},"one_line_summary":"GopherCite produces answers with supporting evidence citations, rated high-quality 80% of the time on Natural Questions and 67% on ELI5, improving to 90% and 80% with abstention on uncertain questions.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A 280 billion parameter model can be trained to answer questions with specific cited evidence from documents and to abstain when uncertain.","strongest_claim":"Our 280 billion parameter model, GopherCite, is able to produce answers with high quality supporting evidence and abstain from answering when unsure. The model's response is found to be high-quality 80% of the time on this Natural Questions subset, and 67% of the time on the ELI5 subset. Abstaining from the third of questions for which it is most unsure improves performance to 90% and 80% respectively.","weakest_assumption":"That human raters' preferences for 'high quality supporting evidence' during RLHP training generalize to produce reliable citations and that the model's internal uncertainty signal for abstention is well-calibrated without introducing new biases."}},"verdict_id":"9612394c-192c-4c1f-a9f9-d2f981f7f4f1"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:728841da86dc62209d64ce03e96a06240b7235c3e889e23851ab60ff245aaa56","target":"record","created_at":"2026-05-17T23:38:14Z","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":"3f27c4af1e562a931e61026b868ea299ed8a7336b9e159d0dbe738400d973823","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-03-21T17:26:29Z","title_canon_sha256":"13c59520008b99b844030d7803e45e1caa24dcb0fc66e41169eb41631e273ec5"},"schema_version":"1.0","source":{"id":"2203.11147","kind":"arxiv","version":1}},"canonical_sha256":"b7416fee07b9e32e6402efbcccf9cc8c75905c39fdbe93293cc1ef5a6e1101d5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b7416fee07b9e32e6402efbcccf9cc8c75905c39fdbe93293cc1ef5a6e1101d5","first_computed_at":"2026-05-17T23:38:14.206266Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:14.206266Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"op1DX+t2vqDDhyd+MoJdrNCcCF3FyRpCPZg7gOtjTgLWJfZofjprS7EdTMsJjC0n6Mz58Vr4mSqN0DpOKl2tCg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:14.206967Z","signed_message":"canonical_sha256_bytes"},"source_id":"2203.11147","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:728841da86dc62209d64ce03e96a06240b7235c3e889e23851ab60ff245aaa56","sha256:440e2e1282e943bb0d459eee5d7ae41d9c6ff75aff0bdf9e54f12cd0f1b5f316"],"state_sha256":"5effc4a5cf19bd6a6005b300ae2952478630e5d3ad75fa541e1e3982991f5ab4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"O9MDJR/kW6jxhrQZUUiMWGFzXk0H+7CLQwrfktSgqzs2J0R0VP0PO5r7uPVjt0IluUmwtvcDE249JtVUKeCqCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-26T22:11:44.320408Z","bundle_sha256":"adfe679079e6c0b21792fe9828110cf668c9d1425a3adfd621018ba1cd00b0b9"}}