{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:LUJO3CWDG3XYAEGXGKBXBQKD3Z","short_pith_number":"pith:LUJO3CWD","schema_version":"1.0","canonical_sha256":"5d12ed8ac336ef8010d7328370c143de6696f20e75738b4699c960991db6d7c4","source":{"kind":"arxiv","id":"2506.12346","version":1},"attestation_state":"computed","paper":{"title":"Refract ICL: Rethinking Example Selection in the Era of Million-Token Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aditi Chaudhary, Arjun R. Akula, Karthik Raman, Kazuma Hashimoto, Krishna Srinivasan, Michael Bendersky","submitted_at":"2025-06-14T04:51:34Z","abstract_excerpt":"The emergence of long-context large language models (LLMs) has enabled the use of hundreds, or even thousands, of demonstrations for in-context learning (ICL) - a previously impractical regime. This paper investigates whether traditional ICL selection strategies, which balance the similarity of ICL examples to the test input (using a text retriever) with diversity within the ICL set, remain effective when utilizing a large number of demonstrations. Our experiments demonstrate that, while longer contexts can accommodate more examples, simply increasing the number of demonstrations does not guar"},"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":"2506.12346","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-14T04:51:34Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"37c75061705c626f7cd197a676b4893d738333a3a2bcf2b217babbf8aaf20454","abstract_canon_sha256":"c391164999d57c8223bdad20983c51d78f9b78b81ba3e7cb5e1ef01d16b96bc0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:21:51.684802Z","signature_b64":"100j6J5HhR2wIeXkfAMY/QuYj/V7pLGX5DSGJrsr1tpgCgVOopBVQXswIV2hJ8S1TYtiybxQ3NaYMxKLAmPMDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5d12ed8ac336ef8010d7328370c143de6696f20e75738b4699c960991db6d7c4","last_reissued_at":"2026-07-05T11:21:51.684294Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:21:51.684294Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Refract ICL: Rethinking Example Selection in the Era of Million-Token Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aditi Chaudhary, Arjun R. Akula, Karthik Raman, Kazuma Hashimoto, Krishna Srinivasan, Michael Bendersky","submitted_at":"2025-06-14T04:51:34Z","abstract_excerpt":"The emergence of long-context large language models (LLMs) has enabled the use of hundreds, or even thousands, of demonstrations for in-context learning (ICL) - a previously impractical regime. This paper investigates whether traditional ICL selection strategies, which balance the similarity of ICL examples to the test input (using a text retriever) with diversity within the ICL set, remain effective when utilizing a large number of demonstrations. Our experiments demonstrate that, while longer contexts can accommodate more examples, simply increasing the number of demonstrations does not guar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.12346","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/2506.12346/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":"2506.12346","created_at":"2026-07-05T11:21:51.684368+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.12346v1","created_at":"2026-07-05T11:21:51.684368+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.12346","created_at":"2026-07-05T11:21:51.684368+00:00"},{"alias_kind":"pith_short_12","alias_value":"LUJO3CWDG3XY","created_at":"2026-07-05T11:21:51.684368+00:00"},{"alias_kind":"pith_short_16","alias_value":"LUJO3CWDG3XYAEGX","created_at":"2026-07-05T11:21:51.684368+00:00"},{"alias_kind":"pith_short_8","alias_value":"LUJO3CWD","created_at":"2026-07-05T11:21:51.684368+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/LUJO3CWDG3XYAEGXGKBXBQKD3Z","json":"https://pith.science/pith/LUJO3CWDG3XYAEGXGKBXBQKD3Z.json","graph_json":"https://pith.science/api/pith-number/LUJO3CWDG3XYAEGXGKBXBQKD3Z/graph.json","events_json":"https://pith.science/api/pith-number/LUJO3CWDG3XYAEGXGKBXBQKD3Z/events.json","paper":"https://pith.science/paper/LUJO3CWD"},"agent_actions":{"view_html":"https://pith.science/pith/LUJO3CWDG3XYAEGXGKBXBQKD3Z","download_json":"https://pith.science/pith/LUJO3CWDG3XYAEGXGKBXBQKD3Z.json","view_paper":"https://pith.science/paper/LUJO3CWD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.12346&json=true","fetch_graph":"https://pith.science/api/pith-number/LUJO3CWDG3XYAEGXGKBXBQKD3Z/graph.json","fetch_events":"https://pith.science/api/pith-number/LUJO3CWDG3XYAEGXGKBXBQKD3Z/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LUJO3CWDG3XYAEGXGKBXBQKD3Z/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LUJO3CWDG3XYAEGXGKBXBQKD3Z/action/storage_attestation","attest_author":"https://pith.science/pith/LUJO3CWDG3XYAEGXGKBXBQKD3Z/action/author_attestation","sign_citation":"https://pith.science/pith/LUJO3CWDG3XYAEGXGKBXBQKD3Z/action/citation_signature","submit_replication":"https://pith.science/pith/LUJO3CWDG3XYAEGXGKBXBQKD3Z/action/replication_record"}},"created_at":"2026-07-05T11:21:51.684368+00:00","updated_at":"2026-07-05T11:21:51.684368+00:00"}