{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QE3U4QFC7QFQCNBLSRSN7YEHWG","short_pith_number":"pith:QE3U4QFC","schema_version":"1.0","canonical_sha256":"81374e40a2fc0b01342b9464dfe087b1a324b2e28ff1adfec8bf42f5288151f2","source":{"kind":"arxiv","id":"2607.01538","version":1},"attestation_state":"computed","paper":{"title":"Can Language Models Actually Retrieve In-Context? Drowning in Documents at Million Token Scale","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Nilesh Gupta, Prasann Singhal, Sewon Min, Siddharth Gollapudi","submitted_at":"2026-07-01T23:38:25Z","abstract_excerpt":"Language models (LMs) raise an intriguing alternative to vector-based retrieval: conditioning on an in-context corpus and directly generating a relevant answer. However, prior work has largely focused on proprietary systems or the smaller-scale reranking task, leaving corpus-scale in-context retrieval largely unexplored. In this work, we present the first systematic study of in-context retrieval on two scales practical retrievers demand: million-token corpora and length-generalization far beyond training-time sizes. We first introduce BlockSearch, a 0.6B LM retriever whose architectural and tr"},"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":"2607.01538","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-01T23:38:25Z","cross_cats_sorted":[],"title_canon_sha256":"60acb3adda6ca3503a557487293b2f10f917cca0696c2a6831cadbb0ae4c8b44","abstract_canon_sha256":"0a5db8744e22c7c02561fdb4b7ba84003c2a89b26b696e67629f27469da97309"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T00:17:03.476429Z","signature_b64":"VO39y0yYhuAken20A/QoNPED2AmfEjgkbqJuLIyz2gllTwgzd0YuoIobmwHlke52LK31SkfP0EX1fwdopsyyBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"81374e40a2fc0b01342b9464dfe087b1a324b2e28ff1adfec8bf42f5288151f2","last_reissued_at":"2026-07-03T00:17:03.475962Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T00:17:03.475962Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Can Language Models Actually Retrieve In-Context? Drowning in Documents at Million Token Scale","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Nilesh Gupta, Prasann Singhal, Sewon Min, Siddharth Gollapudi","submitted_at":"2026-07-01T23:38:25Z","abstract_excerpt":"Language models (LMs) raise an intriguing alternative to vector-based retrieval: conditioning on an in-context corpus and directly generating a relevant answer. However, prior work has largely focused on proprietary systems or the smaller-scale reranking task, leaving corpus-scale in-context retrieval largely unexplored. In this work, we present the first systematic study of in-context retrieval on two scales practical retrievers demand: million-token corpora and length-generalization far beyond training-time sizes. We first introduce BlockSearch, a 0.6B LM retriever whose architectural and tr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.01538","kind":"arxiv","version":1},"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/2607.01538/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":"2607.01538","created_at":"2026-07-03T00:17:03.476021+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.01538v1","created_at":"2026-07-03T00:17:03.476021+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.01538","created_at":"2026-07-03T00:17:03.476021+00:00"},{"alias_kind":"pith_short_12","alias_value":"QE3U4QFC7QFQ","created_at":"2026-07-03T00:17:03.476021+00:00"},{"alias_kind":"pith_short_16","alias_value":"QE3U4QFC7QFQCNBL","created_at":"2026-07-03T00:17:03.476021+00:00"},{"alias_kind":"pith_short_8","alias_value":"QE3U4QFC","created_at":"2026-07-03T00:17:03.476021+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG","json":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG.json","graph_json":"https://pith.science/api/pith-number/QE3U4QFC7QFQCNBLSRSN7YEHWG/graph.json","events_json":"https://pith.science/api/pith-number/QE3U4QFC7QFQCNBLSRSN7YEHWG/events.json","paper":"https://pith.science/paper/QE3U4QFC"},"agent_actions":{"view_html":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG","download_json":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG.json","view_paper":"https://pith.science/paper/QE3U4QFC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.01538&json=true","fetch_graph":"https://pith.science/api/pith-number/QE3U4QFC7QFQCNBLSRSN7YEHWG/graph.json","fetch_events":"https://pith.science/api/pith-number/QE3U4QFC7QFQCNBLSRSN7YEHWG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG/action/storage_attestation","attest_author":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG/action/author_attestation","sign_citation":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG/action/citation_signature","submit_replication":"https://pith.science/pith/QE3U4QFC7QFQCNBLSRSN7YEHWG/action/replication_record"}},"created_at":"2026-07-03T00:17:03.476021+00:00","updated_at":"2026-07-03T00:17:03.476021+00:00"}