{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:SVQYSEI2CVWDM2FRZHTKWIWQFF","short_pith_number":"pith:SVQYSEI2","schema_version":"1.0","canonical_sha256":"956189111a156c3668b1c9e6ab22d0294610edc679802b8f4d28fbf8b5cd50fc","source":{"kind":"arxiv","id":"2305.18466","version":3},"attestation_state":"computed","paper":{"title":"Test-Time Training on Nearest Neighbors for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Moritz Hardt, Yu Sun","submitted_at":"2023-05-29T08:03:28Z","abstract_excerpt":"Many recent efforts augment language models with retrieval, by adding retrieved data to the input context. For this approach to succeed, the retrieved data must be added at both training and test time. Moreover, as input length grows linearly with the size of retrieved data, cost in computation and memory grows quadratically for modern Transformers. To avoid these complications, we simply fine-tune the model on retrieved data at test time, using its standard training setup. We build a large-scale distributed index based on text embeddings of the Pile dataset. For each test input, our system re"},"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":"2305.18466","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-29T08:03:28Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8fd5243bc8882fb7981eac4d362f30a2096e9ce8f7095132c0ba118e2c6a9c71","abstract_canon_sha256":"4c76bb56d2b14821cb032b6ed409e7d2c0c63c093105f8044bfeeb4a5f17d6c6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:40:53.813418Z","signature_b64":"4BI9/Uf4DztB3i9ZsNo9aKfw7gkgQge5xHlZeSdvn2ST2adi2YbYdqPpUo9ns2cv+3QjRGP0ed1/kWpusluoDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"956189111a156c3668b1c9e6ab22d0294610edc679802b8f4d28fbf8b5cd50fc","last_reissued_at":"2026-07-05T07:40:53.812944Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:40:53.812944Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Test-Time Training on Nearest Neighbors for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Moritz Hardt, Yu Sun","submitted_at":"2023-05-29T08:03:28Z","abstract_excerpt":"Many recent efforts augment language models with retrieval, by adding retrieved data to the input context. For this approach to succeed, the retrieved data must be added at both training and test time. Moreover, as input length grows linearly with the size of retrieved data, cost in computation and memory grows quadratically for modern Transformers. To avoid these complications, we simply fine-tune the model on retrieved data at test time, using its standard training setup. We build a large-scale distributed index based on text embeddings of the Pile dataset. For each test input, our system re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.18466","kind":"arxiv","version":3},"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/2305.18466/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":"2305.18466","created_at":"2026-07-05T07:40:53.812999+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.18466v3","created_at":"2026-07-05T07:40:53.812999+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.18466","created_at":"2026-07-05T07:40:53.812999+00:00"},{"alias_kind":"pith_short_12","alias_value":"SVQYSEI2CVWD","created_at":"2026-07-05T07:40:53.812999+00:00"},{"alias_kind":"pith_short_16","alias_value":"SVQYSEI2CVWDM2FR","created_at":"2026-07-05T07:40:53.812999+00:00"},{"alias_kind":"pith_short_8","alias_value":"SVQYSEI2","created_at":"2026-07-05T07:40:53.812999+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26515","citing_title":"Forget, Anticipate and Adapt: Test Time Training for Long Videos","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26515","citing_title":"Forget, Anticipate and Adapt: Test Time Training for Long Videos","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28349","citing_title":"HMARS: A Hierarchical Multi-Agent Memory System for Long-Context Reasoning","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2601.16175","citing_title":"Learning to Discover at Test Time","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13369","citing_title":"Query-Conditioned Test-Time Self-Training for Large Language Models","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2407.04620","citing_title":"Learning to (Learn at Test Time): RNNs with Expressive Hidden States","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13369","citing_title":"Query-Conditioned Test-Time Self-Training for Large Language Models","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11328","citing_title":"Epistemic Uncertainty for Test-Time Discovery","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF","json":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF.json","graph_json":"https://pith.science/api/pith-number/SVQYSEI2CVWDM2FRZHTKWIWQFF/graph.json","events_json":"https://pith.science/api/pith-number/SVQYSEI2CVWDM2FRZHTKWIWQFF/events.json","paper":"https://pith.science/paper/SVQYSEI2"},"agent_actions":{"view_html":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF","download_json":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF.json","view_paper":"https://pith.science/paper/SVQYSEI2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.18466&json=true","fetch_graph":"https://pith.science/api/pith-number/SVQYSEI2CVWDM2FRZHTKWIWQFF/graph.json","fetch_events":"https://pith.science/api/pith-number/SVQYSEI2CVWDM2FRZHTKWIWQFF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF/action/storage_attestation","attest_author":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF/action/author_attestation","sign_citation":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF/action/citation_signature","submit_replication":"https://pith.science/pith/SVQYSEI2CVWDM2FRZHTKWIWQFF/action/replication_record"}},"created_at":"2026-07-05T07:40:53.812999+00:00","updated_at":"2026-07-05T07:40:53.812999+00:00"}