{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:VHW3PIJ57ONRQ6DA2D7WZ7EJWB","short_pith_number":"pith:VHW3PIJ5","schema_version":"1.0","canonical_sha256":"a9edb7a13dfb9b187860d0ff6cfc89b04205eea9087d41afbb755128156a4cca","source":{"kind":"arxiv","id":"2402.18819","version":2},"attestation_state":"computed","paper":{"title":"Dual Operating Modes of In-Context Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Kangwook Lee, Ziqian Lin","submitted_at":"2024-02-29T03:06:10Z","abstract_excerpt":"In-context learning (ICL) exhibits dual operating modes: task learning, i.e., acquiring a new skill from in-context samples, and task retrieval, i.e., locating and activating a relevant pretrained skill. Recent theoretical work investigates various mathematical models to analyze ICL, but existing models explain only one operating mode at a time. We introduce a probabilistic model, with which one can explain the dual operating modes of ICL simultaneously. Focusing on in-context learning of linear functions, we extend existing models for pretraining data by introducing multiple task groups and t"},"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":"2402.18819","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-29T03:06:10Z","cross_cats_sorted":[],"title_canon_sha256":"b90eb675e5afd0b3ccf9d1cca284288df3dce881131433b7adab7fd6dfbb4642","abstract_canon_sha256":"86b8c386223c0711c787fc500fd9177600c133e79f359f9416d762dc77dacdff"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:51:14.028546Z","signature_b64":"iL6TDezJh0TQk86mAXUrvva1KgGz5lbLDnTTSRWraF0NvCJM1ZHnD6BmZ/IYOsBRpO3SaMiD/yu6j/V99ZMgBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a9edb7a13dfb9b187860d0ff6cfc89b04205eea9087d41afbb755128156a4cca","last_reissued_at":"2026-07-05T08:51:14.028011Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:51:14.028011Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dual Operating Modes of In-Context Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Kangwook Lee, Ziqian Lin","submitted_at":"2024-02-29T03:06:10Z","abstract_excerpt":"In-context learning (ICL) exhibits dual operating modes: task learning, i.e., acquiring a new skill from in-context samples, and task retrieval, i.e., locating and activating a relevant pretrained skill. Recent theoretical work investigates various mathematical models to analyze ICL, but existing models explain only one operating mode at a time. We introduce a probabilistic model, with which one can explain the dual operating modes of ICL simultaneously. Focusing on in-context learning of linear functions, we extend existing models for pretraining data by introducing multiple task groups and t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.18819","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/2402.18819/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":"2402.18819","created_at":"2026-07-05T08:51:14.028082+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.18819v2","created_at":"2026-07-05T08:51:14.028082+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.18819","created_at":"2026-07-05T08:51:14.028082+00:00"},{"alias_kind":"pith_short_12","alias_value":"VHW3PIJ57ONR","created_at":"2026-07-05T08:51:14.028082+00:00"},{"alias_kind":"pith_short_16","alias_value":"VHW3PIJ57ONRQ6DA","created_at":"2026-07-05T08:51:14.028082+00:00"},{"alias_kind":"pith_short_8","alias_value":"VHW3PIJ5","created_at":"2026-07-05T08:51:14.028082+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.16217","citing_title":"Towards Compute-Optimal Many-Shot In-Context Learning","ref_index":2002,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB","json":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB.json","graph_json":"https://pith.science/api/pith-number/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/graph.json","events_json":"https://pith.science/api/pith-number/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/events.json","paper":"https://pith.science/paper/VHW3PIJ5"},"agent_actions":{"view_html":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB","download_json":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB.json","view_paper":"https://pith.science/paper/VHW3PIJ5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.18819&json=true","fetch_graph":"https://pith.science/api/pith-number/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/graph.json","fetch_events":"https://pith.science/api/pith-number/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/action/storage_attestation","attest_author":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/action/author_attestation","sign_citation":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/action/citation_signature","submit_replication":"https://pith.science/pith/VHW3PIJ57ONRQ6DA2D7WZ7EJWB/action/replication_record"}},"created_at":"2026-07-05T08:51:14.028082+00:00","updated_at":"2026-07-05T08:51:14.028082+00:00"}