{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:PX63XKQTP4IMHWPZBFOIQNBAWT","short_pith_number":"pith:PX63XKQT","schema_version":"1.0","canonical_sha256":"7dfdbbaa137f10c3d9f9095c883420b4f20350a565ad7f7b1f7e0a3c077fb997","source":{"kind":"arxiv","id":"2301.02828","version":2},"attestation_state":"computed","paper":{"title":"Why do Nearest Neighbor Language Models Work?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Frank F. Xu, Graham Neubig, Uri Alon","submitted_at":"2023-01-07T11:12:36Z","abstract_excerpt":"Language models (LMs) compute the probability of a text by sequentially computing a representation of an already-seen context and using this representation to predict the next word. Currently, most LMs calculate these representations through a neural network consuming the immediate previous context. However recently, retrieval-augmented LMs have shown to improve over standard neural LMs, by accessing information retrieved from a large datastore, in addition to their standard, parametric, next-word prediction. In this paper, we set out to understand why retrieval-augmented language models, and "},"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":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2301.02828","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-01-07T11:12:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"c862f30af65072b636e79c403fc5c7eba77e811273322d513ca21f87bc340c2e","abstract_canon_sha256":"7ac2096bc0d0c3ae518c2fd501ced877ececcb467c55021aef376e6e9704d7c7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:33:23.482672Z","signature_b64":"RYPcRmlP+vQWP+h+OJrQp9YWeUGBBSELP6U/Uj/CytIUN+woPjOzKgsCidCxxKDW5n9oUNiLBE32vgioC7VjDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7dfdbbaa137f10c3d9f9095c883420b4f20350a565ad7f7b1f7e0a3c077fb997","last_reissued_at":"2026-07-05T05:33:23.482258Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:33:23.482258Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Why do Nearest Neighbor Language Models Work?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Frank F. Xu, Graham Neubig, Uri Alon","submitted_at":"2023-01-07T11:12:36Z","abstract_excerpt":"Language models (LMs) compute the probability of a text by sequentially computing a representation of an already-seen context and using this representation to predict the next word. Currently, most LMs calculate these representations through a neural network consuming the immediate previous context. However recently, retrieval-augmented LMs have shown to improve over standard neural LMs, by accessing information retrieved from a large datastore, in addition to their standard, parametric, next-word prediction. In this paper, we set out to understand why retrieval-augmented language models, and "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2301.02828","kind":"arxiv","version":2},"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/2301.02828/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2301.02828","created_at":"2026-07-05T05:33:23.482320+00:00"},{"alias_kind":"arxiv_version","alias_value":"2301.02828v2","created_at":"2026-07-05T05:33:23.482320+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2301.02828","created_at":"2026-07-05T05:33:23.482320+00:00"},{"alias_kind":"pith_short_12","alias_value":"PX63XKQTP4IM","created_at":"2026-07-05T05:33:23.482320+00:00"},{"alias_kind":"pith_short_16","alias_value":"PX63XKQTP4IMHWPZ","created_at":"2026-07-05T05:33:23.482320+00:00"},{"alias_kind":"pith_short_8","alias_value":"PX63XKQT","created_at":"2026-07-05T05:33:23.482320+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.14006","citing_title":"Experimental comparison of graph-based approximate nearest neighbor search algorithms on edge devices","ref_index":5,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT","json":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT.json","graph_json":"https://pith.science/api/pith-number/PX63XKQTP4IMHWPZBFOIQNBAWT/graph.json","events_json":"https://pith.science/api/pith-number/PX63XKQTP4IMHWPZBFOIQNBAWT/events.json","paper":"https://pith.science/paper/PX63XKQT"},"agent_actions":{"view_html":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT","download_json":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT.json","view_paper":"https://pith.science/paper/PX63XKQT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2301.02828&json=true","fetch_graph":"https://pith.science/api/pith-number/PX63XKQTP4IMHWPZBFOIQNBAWT/graph.json","fetch_events":"https://pith.science/api/pith-number/PX63XKQTP4IMHWPZBFOIQNBAWT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT/action/storage_attestation","attest_author":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT/action/author_attestation","sign_citation":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT/action/citation_signature","submit_replication":"https://pith.science/pith/PX63XKQTP4IMHWPZBFOIQNBAWT/action/replication_record"}},"created_at":"2026-07-05T05:33:23.482320+00:00","updated_at":"2026-07-05T05:33:23.482320+00:00"}