{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ESZYU3G4Q5ZLKCUID624XZXTOE","short_pith_number":"pith:ESZYU3G4","schema_version":"1.0","canonical_sha256":"24b38a6cdc8772b50a881fb5cbe6f37129ef5a0837fdf323c6051186a050f50e","source":{"kind":"arxiv","id":"2401.06469","version":3},"attestation_state":"computed","paper":{"title":"Batch-ICL: Effective, Efficient, and Order-Agnostic In-Context Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Ang Lv, Hansen Ha, Kaiyi Zhang, Rui Yan, Tao Xu, Yuhan Chen","submitted_at":"2024-01-12T09:31:17Z","abstract_excerpt":"In this paper, by treating in-context learning (ICL) as a meta-optimization process, we explain why LLMs are sensitive to the order of ICL examples. This understanding leads us to the development of Batch-ICL, an effective, efficient, and order-agnostic inference algorithm for ICL. Differing from the standard N-shot learning approach, Batch-ICL employs $N$ separate 1-shot forward computations and aggregates the resulting meta-gradients. These aggregated meta-gradients are then applied to the forward computation of a zero-shot query to generate the final prediction. This batch processing approa"},"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":"2401.06469","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-01-12T09:31:17Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"83ddf61445caf162ba5031170fe415df9e0d35410963672a447764f915864c98","abstract_canon_sha256":"b4c7dae5a2d3e630f68dfd204d485acb70a0248bec050aa091fe82c49b7b9c88"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:27:45.378203Z","signature_b64":"Kr76HVWS+6A7GKuLhh3Me//i9lIaiB2LzMkzMR3P6DnLAvUP7hLPgVJS0PIhLk7xo1TvQKyfey+dmp1+qhY9Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"24b38a6cdc8772b50a881fb5cbe6f37129ef5a0837fdf323c6051186a050f50e","last_reissued_at":"2026-07-05T08:27:45.377705Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:27:45.377705Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Batch-ICL: Effective, Efficient, and Order-Agnostic In-Context Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Ang Lv, Hansen Ha, Kaiyi Zhang, Rui Yan, Tao Xu, Yuhan Chen","submitted_at":"2024-01-12T09:31:17Z","abstract_excerpt":"In this paper, by treating in-context learning (ICL) as a meta-optimization process, we explain why LLMs are sensitive to the order of ICL examples. This understanding leads us to the development of Batch-ICL, an effective, efficient, and order-agnostic inference algorithm for ICL. Differing from the standard N-shot learning approach, Batch-ICL employs $N$ separate 1-shot forward computations and aggregates the resulting meta-gradients. These aggregated meta-gradients are then applied to the forward computation of a zero-shot query to generate the final prediction. This batch processing approa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.06469","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/2401.06469/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":"2401.06469","created_at":"2026-07-05T08:27:45.377767+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.06469v3","created_at":"2026-07-05T08:27:45.377767+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.06469","created_at":"2026-07-05T08:27:45.377767+00:00"},{"alias_kind":"pith_short_12","alias_value":"ESZYU3G4Q5ZL","created_at":"2026-07-05T08:27:45.377767+00:00"},{"alias_kind":"pith_short_16","alias_value":"ESZYU3G4Q5ZLKCUI","created_at":"2026-07-05T08:27:45.377767+00:00"},{"alias_kind":"pith_short_8","alias_value":"ESZYU3G4","created_at":"2026-07-05T08:27:45.377767+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.12004","citing_title":"Improving Data and Parameter Efficiency of Neural Language Models Using Representation Analysis","ref_index":139,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE","json":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE.json","graph_json":"https://pith.science/api/pith-number/ESZYU3G4Q5ZLKCUID624XZXTOE/graph.json","events_json":"https://pith.science/api/pith-number/ESZYU3G4Q5ZLKCUID624XZXTOE/events.json","paper":"https://pith.science/paper/ESZYU3G4"},"agent_actions":{"view_html":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE","download_json":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE.json","view_paper":"https://pith.science/paper/ESZYU3G4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.06469&json=true","fetch_graph":"https://pith.science/api/pith-number/ESZYU3G4Q5ZLKCUID624XZXTOE/graph.json","fetch_events":"https://pith.science/api/pith-number/ESZYU3G4Q5ZLKCUID624XZXTOE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE/action/storage_attestation","attest_author":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE/action/author_attestation","sign_citation":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE/action/citation_signature","submit_replication":"https://pith.science/pith/ESZYU3G4Q5ZLKCUID624XZXTOE/action/replication_record"}},"created_at":"2026-07-05T08:27:45.377767+00:00","updated_at":"2026-07-05T08:27:45.377767+00:00"}