{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:6DVG7MV3CDNLUNQDZKFYQHEVL4","short_pith_number":"pith:6DVG7MV3","schema_version":"1.0","canonical_sha256":"f0ea6fb2bb10daba3603ca8b881c955f1db39e171f12f9fa973aed36e815f0c0","source":{"kind":"arxiv","id":"1909.01066","version":2},"attestation_state":"computed","paper":{"title":"Language Models as Knowledge Bases?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Pretrained language models already contain relational knowledge that can be accessed through fill-in-the-blank queries without any fine-tuning.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexander H. Miller, Anton Bakhtin, Fabio Petroni, Patrick Lewis, Sebastian Riedel, Tim Rockt\\\"aschel, Yuxiang WU","submitted_at":"2019-09-03T11:11:08Z","abstract_excerpt":"Recent progress in pretraining language models on large textual corpora led to a surge of improvements for downstream NLP tasks. Whilst learning linguistic knowledge, these models may also be storing relational knowledge present in the training data, and may be able to answer queries structured as \"fill-in-the-blank\" cloze statements. Language models have many advantages over structured knowledge bases: they require no schema engineering, allow practitioners to query about an open class of relations, are easy to extend to more data, and require no human supervision to train. We present an in-d"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":true,"formal_links_present":true},"canonical_record":{"source":{"id":"1909.01066","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-09-03T11:11:08Z","cross_cats_sorted":[],"title_canon_sha256":"806066cd870d90a814cc6da54f47e7c2bdbe38284f76cd67d7056b51a66920b5","abstract_canon_sha256":"b4b9d25078db0085f26c4e02c555f1804072205847706550f402b5812b96137f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:48.129448Z","signature_b64":"2j2ui7JULz3WS1tBj61L5kRkcQjDIeHSFhJ1nQIQxLVx1l7nfJsxxulDuiCnDwB6joa2Yw24PpNZxye9UMvVBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f0ea6fb2bb10daba3603ca8b881c955f1db39e171f12f9fa973aed36e815f0c0","last_reissued_at":"2026-05-17T23:38:48.128807Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:48.128807Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language Models as Knowledge Bases?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Pretrained language models already contain relational knowledge that can be accessed through fill-in-the-blank queries without any fine-tuning.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexander H. Miller, Anton Bakhtin, Fabio Petroni, Patrick Lewis, Sebastian Riedel, Tim Rockt\\\"aschel, Yuxiang WU","submitted_at":"2019-09-03T11:11:08Z","abstract_excerpt":"Recent progress in pretraining language models on large textual corpora led to a surge of improvements for downstream NLP tasks. Whilst learning linguistic knowledge, these models may also be storing relational knowledge present in the training data, and may be able to answer queries structured as \"fill-in-the-blank\" cloze statements. Language models have many advantages over structured knowledge bases: they require no schema engineering, allow practitioners to query about an open class of relations, are easy to extend to more data, and require no human supervision to train. We present an in-d"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"without fine-tuning, BERT contains relational knowledge competitive with traditional NLP methods that have some access to oracle knowledge","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That cloze-statement accuracy directly measures stored factual knowledge rather than surface-level statistical patterns or memorization of training co-occurrences.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"BERT stores relational knowledge extractable via cloze queries without fine-tuning and matches supervised baselines on open-domain QA tasks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Pretrained language models already contain relational knowledge that can be accessed through fill-in-the-blank queries without any fine-tuning.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f2b098dc9b1a9b97d49f85da85b0186698d7cede9a9132b2da521803ed7dc70e"},"source":{"id":"1909.01066","kind":"arxiv","version":2},"verdict":{"id":"66447a3c-f6d0-4078-aa9d-527610dd88ea","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T10:48:00.004958Z","strongest_claim":"without fine-tuning, BERT contains relational knowledge competitive with traditional NLP methods that have some access to oracle knowledge","one_line_summary":"BERT stores relational knowledge extractable via cloze queries without fine-tuning and matches supervised baselines on open-domain QA tasks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That cloze-statement accuracy directly measures stored factual knowledge rather than surface-level statistical patterns or memorization of training co-occurrences.","pith_extraction_headline":"Pretrained language models already contain relational knowledge that can be accessed through fill-in-the-blank queries without any fine-tuning."},"references":{"count":300,"sample":[{"doi":"","year":null,"title":"Overview of the","work_id":"d1345a3f-9d55-4e0a-99f6-2bd1f4513f3a","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1901,"title":"Learning and evaluating general linguistic intelligence","work_id":"29fd3485-ad71-4ae6-b96f-d35675eddb26","ref_index":2,"cited_arxiv_id":"1901.11373","is_internal_anchor":true},{"doi":"","year":2018,"title":"Lample, Guillaume and Ott, Myle and Conneau, Alexis and Denoyer, Ludovic and Ranzato, Marc'Aurelio , month = nov, year =. Phrase-. Proceedings of the 2018","work_id":"027e5bc2-c300-414c-b94b-eb3006d73004","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"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":"Transactions of the Association of Computational Linguistics , author =","work_id":"c4cfe119-cd35-4064-8a33-96827a6e4234","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":300,"snapshot_sha256":"a71332a8631d8487b57fd8293dff1ab6091f6599c86ae1dc1374e8815b147dfa","internal_anchors":41},"formal_canon":{"evidence_count":2,"snapshot_sha256":"8a6297db481a34523fc2f8ca7b195f002fec9bb7f63762ab00149f9834872a68"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"1909.01066","created_at":"2026-05-17T23:38:48.128917+00:00"},{"alias_kind":"arxiv_version","alias_value":"1909.01066v2","created_at":"2026-05-17T23:38:48.128917+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1909.01066","created_at":"2026-05-17T23:38:48.128917+00:00"},{"alias_kind":"pith_short_12","alias_value":"6DVG7MV3CDNL","created_at":"2026-05-18T12:33:10.108867+00:00"},{"alias_kind":"pith_short_16","alias_value":"6DVG7MV3CDNLUNQD","created_at":"2026-05-18T12:33:10.108867+00:00"},{"alias_kind":"pith_short_8","alias_value":"6DVG7MV3","created_at":"2026-05-18T12:33:10.108867+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":24,"internal_anchor_count":24,"sample":[{"citing_arxiv_id":"2401.00761","citing_title":"Identifying the Achilles' Heel: An Iterative Method for Dynamically Uncovering Factual Errors in Large Language Models","ref_index":56,"is_internal_anchor":true},{"citing_arxiv_id":"2502.11008","citing_title":"CounterBench: Evaluating and Improving Counterfactual Reasoning in Large Language Models","ref_index":34,"is_internal_anchor":true},{"citing_arxiv_id":"2509.08461","citing_title":"Adapting Vision-Language Models for Neutrino Event Classification in High-Energy Physics","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2510.17934","citing_title":"AtlasKV: Augmenting LLMs with Billion-Scale Knowledge Graphs in 20GB VRAM","ref_index":25,"is_internal_anchor":true},{"citing_arxiv_id":"2511.02626","citing_title":"Understanding New-Knowledge-Induced Factual Hallucinations in LLMs: Analysis and Interpretation","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2309.11495","citing_title":"Chain-of-Verification Reduces Hallucination in Large Language Models","ref_index":162,"is_internal_anchor":true},{"citing_arxiv_id":"2204.00598","citing_title":"Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language","ref_index":116,"is_internal_anchor":true},{"citing_arxiv_id":"2604.06179","citing_title":"ARIA: Adaptive Retrieval Intelligence Assistant -- A Multimodal RAG Framework for Domain-Specific Engineering Education","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2401.18059","citing_title":"RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval","ref_index":131,"is_internal_anchor":true},{"citing_arxiv_id":"2002.08909","citing_title":"REALM: Retrieval-Augmented Language Model Pre-Training","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2205.00445","citing_title":"MRKL Systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14169","citing_title":"BOOKMARKS: Efficient Active Storyline Memory for Role-playing","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2002.08910","citing_title":"How Much Knowledge Can You Pack Into the Parameters of a Language Model?","ref_index":64,"is_internal_anchor":true},{"citing_arxiv_id":"2604.14172","citing_title":"Tug-of-War within A Decade: Conflict Resolution in Vulnerability Analysis via Teacher-Guided Retrieval-Augmented Generations","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2002.08155","citing_title":"CodeBERT: A Pre-Trained Model for Programming and Natural Languages","ref_index":51,"is_internal_anchor":true},{"citing_arxiv_id":"2502.02737","citing_title":"SmolLM2: When Smol Goes Big -- Data-Centric Training of a Small Language Model","ref_index":210,"is_internal_anchor":true},{"citing_arxiv_id":"2310.11511","citing_title":"Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection","ref_index":88,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10640","citing_title":"Towards Understanding Continual Factual Knowledge Acquisition of Language Models: From Theory to Algorithm","ref_index":45,"is_internal_anchor":true},{"citing_arxiv_id":"2207.05608","citing_title":"Inner Monologue: Embodied Reasoning through Planning with Language Models","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2605.05459","citing_title":"Privacy Without Losing Place: A Paradigm for Private Retrieval in Spatial RAGs","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.01359","citing_title":"Structural Ranking of the Cognitive Plausibility of Computational Models of Analogy and Metaphors with the Minimal Cognitive Grid","ref_index":44,"is_internal_anchor":true},{"citing_arxiv_id":"2604.05732","citing_title":"Graph Topology Information Enhanced Heterogeneous Graph Representation Learning","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2604.18124","citing_title":"TLoRA: Task-aware Low Rank Adaptation of Large Language Models","ref_index":49,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15945","citing_title":"RAGognizer: Hallucination-Aware Fine-Tuning via Detection Head Integration","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4","json":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4.json","graph_json":"https://pith.science/api/pith-number/6DVG7MV3CDNLUNQDZKFYQHEVL4/graph.json","events_json":"https://pith.science/api/pith-number/6DVG7MV3CDNLUNQDZKFYQHEVL4/events.json","paper":"https://pith.science/paper/6DVG7MV3"},"agent_actions":{"view_html":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4","download_json":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4.json","view_paper":"https://pith.science/paper/6DVG7MV3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1909.01066&json=true","fetch_graph":"https://pith.science/api/pith-number/6DVG7MV3CDNLUNQDZKFYQHEVL4/graph.json","fetch_events":"https://pith.science/api/pith-number/6DVG7MV3CDNLUNQDZKFYQHEVL4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4/action/storage_attestation","attest_author":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4/action/author_attestation","sign_citation":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4/action/citation_signature","submit_replication":"https://pith.science/pith/6DVG7MV3CDNLUNQDZKFYQHEVL4/action/replication_record"}},"created_at":"2026-05-17T23:38:48.128917+00:00","updated_at":"2026-05-17T23:38:48.128917+00:00"}