{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2020:TOWXDDJWQSC2EPTIXWHQYSZV5P","short_pith_number":"pith:TOWXDDJW","canonical_record":{"source":{"id":"2002.08909","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-10T18:40:59Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4cc1d8fa32eb3843fc491ea769947edd01955ffa807b5f78126a614095724e51","abstract_canon_sha256":"a232dfc972dce5d964abce327735d56ef564c47b6287fe4dc0ed536c127173cb"},"schema_version":"1.0"},"canonical_sha256":"9bad718d368485a23e68bd8f0c4b35ebf7fe612a69c5579e3ba4838f495dc45e","source":{"kind":"arxiv","id":"2002.08909","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2002.08909","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"arxiv_version","alias_value":"2002.08909v1","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.08909","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"pith_short_12","alias_value":"TOWXDDJWQSC2","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"TOWXDDJWQSC2EPTI","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"TOWXDDJW","created_at":"2026-05-18T12:33:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2020:TOWXDDJWQSC2EPTIXWHQYSZV5P","target":"record","payload":{"canonical_record":{"source":{"id":"2002.08909","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-10T18:40:59Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4cc1d8fa32eb3843fc491ea769947edd01955ffa807b5f78126a614095724e51","abstract_canon_sha256":"a232dfc972dce5d964abce327735d56ef564c47b6287fe4dc0ed536c127173cb"},"schema_version":"1.0"},"canonical_sha256":"9bad718d368485a23e68bd8f0c4b35ebf7fe612a69c5579e3ba4838f495dc45e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:52.830061Z","signature_b64":"QyWzjAl+tc7l6pvFXoGlXs8eGo/gUbsVDYcY0kkKMf4yaMDEdzt2R2xuMgj2uW+w4JHd5hG9pcx14jEuYcnODQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9bad718d368485a23e68bd8f0c4b35ebf7fe612a69c5579e3ba4838f495dc45e","last_reissued_at":"2026-05-17T23:38:52.829438Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:52.829438Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2002.08909","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":"2026-05-17T23:38:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"BUfUgxI6VHbxKiSyLtstoL4WQXu/xeN0Q3LMsag080xfFxdF1Ppap6yILaLBR8cc+/3y/sdFqjC+fvQ26UpICA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T12:05:50.178525Z"},"content_sha256":"ff3948f030e7b895c0b0b170b34415d9ad6761bc0aff665b35d9289f40311b7d","schema_version":"1.0","event_id":"sha256:ff3948f030e7b895c0b0b170b34415d9ad6761bc0aff665b35d9289f40311b7d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2020:TOWXDDJWQSC2EPTIXWHQYSZV5P","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"REALM: Retrieval-Augmented Language Model Pre-Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Language models pre-trained with an integrated retriever over a document corpus outperform prior methods on open-domain question answering by 4 to 16 percent.","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Kelvin Guu, Kenton Lee, Ming-Wei Chang, Panupong Pasupat, Zora Tung","submitted_at":"2020-02-10T18:40:59Z","abstract_excerpt":"Language model pre-training has been shown to capture a surprising amount of world knowledge, crucial for NLP tasks such as question answering. However, this knowledge is stored implicitly in the parameters of a neural network, requiring ever-larger networks to cover more facts.\n  To capture knowledge in a more modular and interpretable way, we augment language model pre-training with a latent knowledge retriever, which allows the model to retrieve and attend over documents from a large corpus such as Wikipedia, used during pre-training, fine-tuning and inference. For the first time, we show h"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We demonstrate the effectiveness of Retrieval-Augmented Language Model pre-training (REALM) by fine-tuning on the challenging task of Open-domain Question Answering (Open-QA). We compare against state-of-the-art models for both explicit and implicit knowledge storage on three popular Open-QA benchmarks, and find that we outperform all previous methods by a significant margin (4-16% absolute accuracy).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That back-propagation through a retrieval step over millions of documents is numerically stable and provides a useful unsupervised learning signal for the retriever parameters.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"REALM augments language-model pre-training with an unsupervised retriever over Wikipedia documents and reports 4-16% absolute gains on open-domain QA benchmarks over prior implicit and explicit knowledge methods.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Language models pre-trained with an integrated retriever over a document corpus outperform prior methods on open-domain question answering by 4 to 16 percent.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"fd05d3e400545758fdf5d7a79da9bff47de60a2fdeaaa7ce976fb7785f375b58"},"source":{"id":"2002.08909","kind":"arxiv","version":1},"verdict":{"id":"0496c68e-e097-4692-837a-27d9f626107a","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T09:52:42.388463Z","strongest_claim":"We demonstrate the effectiveness of Retrieval-Augmented Language Model pre-training (REALM) by fine-tuning on the challenging task of Open-domain Question Answering (Open-QA). We compare against state-of-the-art models for both explicit and implicit knowledge storage on three popular Open-QA benchmarks, and find that we outperform all previous methods by a significant margin (4-16% absolute accuracy).","one_line_summary":"REALM augments language-model pre-training with an unsupervised retriever over Wikipedia documents and reports 4-16% absolute gains on open-domain QA benchmarks over prior implicit and explicit knowledge methods.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That back-propagation through a retrieval step over millions of documents is numerically stable and provides a useful unsupervised learning signal for the retriever parameters.","pith_extraction_headline":"Language models pre-trained with an integrated retriever over a document corpus outperform prior methods on open-domain question answering by 4 to 16 percent."},"references":{"count":20,"sample":[{"doi":"","year":1911,"title":"arXiv preprint arXiv:1911.10470 , year=","work_id":"bba0e5b4-dccd-4c07-86fb-2dae46efdce9","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Neural Machine Translation by Jointly Learning to Align and Translate","work_id":"d831e763-d530-4029-a65c-ac595d82cb2a","ref_index":2,"cited_arxiv_id":"1409.0473","is_internal_anchor":true},{"doi":"","year":2013,"title":"Semantic parsing on freebase from question-answer pairs","work_id":"55a52a96-0321-4fc0-a8f5-eccec466bc2d","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding","work_id":"ed240a10-5b19-406c-baa5-30803f465785","ref_index":4,"cited_arxiv_id":"1810.04805","is_internal_anchor":true},{"doi":"","year":null,"title":"Neural Turing Machines","work_id":"0a5ce53c-9670-42b9-8be5-386de7eed50c","ref_index":5,"cited_arxiv_id":"1410.5401","is_internal_anchor":true}],"resolved_work":20,"snapshot_sha256":"217ebf269247a333e779ed5dbbf87e45e2bf0a2c4d5f6cbb1950fc955a6166ce","internal_anchors":11},"formal_canon":{"evidence_count":3,"snapshot_sha256":"921de4a55a8393b7609ce9eec082d809fbe45062a568b445ccd3f0a7c006572d"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"0496c68e-e097-4692-837a-27d9f626107a"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"S6bvVHGN494p6ifiG8Ag7Rmgf+MzJEP+6XjLlkAHioLIO24S73cihDyVIDGqvquqM1vL/aqXGLlZZMjIxlhDDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T12:05:50.179033Z"},"content_sha256":"aef2b7da48cd080d1c233d59ed2ed42c1dea9be8711657ac9f84c079b9938614","schema_version":"1.0","event_id":"sha256:aef2b7da48cd080d1c233d59ed2ed42c1dea9be8711657ac9f84c079b9938614"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/TOWXDDJWQSC2EPTIXWHQYSZV5P/bundle.json","state_url":"https://pith.science/pith/TOWXDDJWQSC2EPTIXWHQYSZV5P/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/TOWXDDJWQSC2EPTIXWHQYSZV5P/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":"2026-05-30T12:05:50Z","links":{"resolver":"https://pith.science/pith/TOWXDDJWQSC2EPTIXWHQYSZV5P","bundle":"https://pith.science/pith/TOWXDDJWQSC2EPTIXWHQYSZV5P/bundle.json","state":"https://pith.science/pith/TOWXDDJWQSC2EPTIXWHQYSZV5P/state.json","well_known_bundle":"https://pith.science/.well-known/pith/TOWXDDJWQSC2EPTIXWHQYSZV5P/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:TOWXDDJWQSC2EPTIXWHQYSZV5P","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":"a232dfc972dce5d964abce327735d56ef564c47b6287fe4dc0ed536c127173cb","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-10T18:40:59Z","title_canon_sha256":"4cc1d8fa32eb3843fc491ea769947edd01955ffa807b5f78126a614095724e51"},"schema_version":"1.0","source":{"id":"2002.08909","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2002.08909","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"arxiv_version","alias_value":"2002.08909v1","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.08909","created_at":"2026-05-17T23:38:52Z"},{"alias_kind":"pith_short_12","alias_value":"TOWXDDJWQSC2","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"TOWXDDJWQSC2EPTI","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"TOWXDDJW","created_at":"2026-05-18T12:33:33Z"}],"graph_snapshots":[{"event_id":"sha256:aef2b7da48cd080d1c233d59ed2ed42c1dea9be8711657ac9f84c079b9938614","target":"graph","created_at":"2026-05-17T23:38:52Z","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":"We demonstrate the effectiveness of Retrieval-Augmented Language Model pre-training (REALM) by fine-tuning on the challenging task of Open-domain Question Answering (Open-QA). We compare against state-of-the-art models for both explicit and implicit knowledge storage on three popular Open-QA benchmarks, and find that we outperform all previous methods by a significant margin (4-16% absolute accuracy)."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That back-propagation through a retrieval step over millions of documents is numerically stable and provides a useful unsupervised learning signal for the retriever parameters."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"REALM augments language-model pre-training with an unsupervised retriever over Wikipedia documents and reports 4-16% absolute gains on open-domain QA benchmarks over prior implicit and explicit knowledge methods."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Language models pre-trained with an integrated retriever over a document corpus outperform prior methods on open-domain question answering by 4 to 16 percent."}],"snapshot_sha256":"fd05d3e400545758fdf5d7a79da9bff47de60a2fdeaaa7ce976fb7785f375b58"},"formal_canon":{"evidence_count":3,"snapshot_sha256":"921de4a55a8393b7609ce9eec082d809fbe45062a568b445ccd3f0a7c006572d"},"paper":{"abstract_excerpt":"Language model pre-training has been shown to capture a surprising amount of world knowledge, crucial for NLP tasks such as question answering. However, this knowledge is stored implicitly in the parameters of a neural network, requiring ever-larger networks to cover more facts.\n  To capture knowledge in a more modular and interpretable way, we augment language model pre-training with a latent knowledge retriever, which allows the model to retrieve and attend over documents from a large corpus such as Wikipedia, used during pre-training, fine-tuning and inference. For the first time, we show h","authors_text":"Kelvin Guu, Kenton Lee, Ming-Wei Chang, Panupong Pasupat, Zora Tung","cross_cats":["cs.LG"],"headline":"Language models pre-trained with an integrated retriever over a document corpus outperform prior methods on open-domain question answering by 4 to 16 percent.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-10T18:40:59Z","title":"REALM: Retrieval-Augmented Language Model Pre-Training"},"references":{"count":20,"internal_anchors":11,"resolved_work":20,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"arXiv preprint arXiv:1911.10470 , year=","work_id":"bba0e5b4-dccd-4c07-86fb-2dae46efdce9","year":1911},{"cited_arxiv_id":"1409.0473","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Neural Machine Translation by Jointly Learning to Align and Translate","work_id":"d831e763-d530-4029-a65c-ac595d82cb2a","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Semantic parsing on freebase from question-answer pairs","work_id":"55a52a96-0321-4fc0-a8f5-eccec466bc2d","year":2013},{"cited_arxiv_id":"1810.04805","doi":"","is_internal_anchor":true,"ref_index":4,"title":"BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding","work_id":"ed240a10-5b19-406c-baa5-30803f465785","year":null},{"cited_arxiv_id":"1410.5401","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Neural Turing Machines","work_id":"0a5ce53c-9670-42b9-8be5-386de7eed50c","year":null}],"snapshot_sha256":"217ebf269247a333e779ed5dbbf87e45e2bf0a2c4d5f6cbb1950fc955a6166ce"},"source":{"id":"2002.08909","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T09:52:42.388463Z","id":"0496c68e-e097-4692-837a-27d9f626107a","model_set":{"reader":"grok-4.3"},"one_line_summary":"REALM augments language-model pre-training with an unsupervised retriever over Wikipedia documents and reports 4-16% absolute gains on open-domain QA benchmarks over prior implicit and explicit knowledge methods.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Language models pre-trained with an integrated retriever over a document corpus outperform prior methods on open-domain question answering by 4 to 16 percent.","strongest_claim":"We demonstrate the effectiveness of Retrieval-Augmented Language Model pre-training (REALM) by fine-tuning on the challenging task of Open-domain Question Answering (Open-QA). We compare against state-of-the-art models for both explicit and implicit knowledge storage on three popular Open-QA benchmarks, and find that we outperform all previous methods by a significant margin (4-16% absolute accuracy).","weakest_assumption":"That back-propagation through a retrieval step over millions of documents is numerically stable and provides a useful unsupervised learning signal for the retriever parameters."}},"verdict_id":"0496c68e-e097-4692-837a-27d9f626107a"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:ff3948f030e7b895c0b0b170b34415d9ad6761bc0aff665b35d9289f40311b7d","target":"record","created_at":"2026-05-17T23:38:52Z","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":"a232dfc972dce5d964abce327735d56ef564c47b6287fe4dc0ed536c127173cb","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-10T18:40:59Z","title_canon_sha256":"4cc1d8fa32eb3843fc491ea769947edd01955ffa807b5f78126a614095724e51"},"schema_version":"1.0","source":{"id":"2002.08909","kind":"arxiv","version":1}},"canonical_sha256":"9bad718d368485a23e68bd8f0c4b35ebf7fe612a69c5579e3ba4838f495dc45e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9bad718d368485a23e68bd8f0c4b35ebf7fe612a69c5579e3ba4838f495dc45e","first_computed_at":"2026-05-17T23:38:52.829438Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:52.829438Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"QyWzjAl+tc7l6pvFXoGlXs8eGo/gUbsVDYcY0kkKMf4yaMDEdzt2R2xuMgj2uW+w4JHd5hG9pcx14jEuYcnODQ==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:52.830061Z","signed_message":"canonical_sha256_bytes"},"source_id":"2002.08909","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ff3948f030e7b895c0b0b170b34415d9ad6761bc0aff665b35d9289f40311b7d","sha256:aef2b7da48cd080d1c233d59ed2ed42c1dea9be8711657ac9f84c079b9938614"],"state_sha256":"fd61b415fad5e6c0529347fbd1eda77b34dd543f01751319d7ae99ad5394cfe8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Fss/PAksD9aDy8ZGuuK7OkUwsZwGIU2lIY8O5g/9eBxnbpoev0X+1gk+RzRC91A2L8O6dpumEFCmsTMiHQ/XCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-30T12:05:50.181373Z","bundle_sha256":"057fb15d8e81e3cf47acc35c63d86613bfc969dba8162048fac0731027c26edf"}}