{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:UTKAP5C2IU6O46HGCWZKLSET7A","short_pith_number":"pith:UTKAP5C2","schema_version":"1.0","canonical_sha256":"a4d407f45a453cee78e615b2a5c893f8195d4678ccb1710f29b0b0ab9ff52baa","source":{"kind":"arxiv","id":"2502.00330","version":1},"attestation_state":"computed","paper":{"title":"From Few to Many: Self-Improving Many-Shot Reasoners Through Iterative Optimization and Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Han Zhou, Hootan Nakhost, Ke Jiang, Ruoxi Sun, Sercan \\\"O. Ar{\\i}k, Xingchen Wan","submitted_at":"2025-02-01T06:23:24Z","abstract_excerpt":"Recent advances in long-context large language models (LLMs) have led to the emerging paradigm of many-shot in-context learning (ICL), where it is observed that scaling many more demonstrating examples beyond the conventional few-shot setup in the context can lead to performance benefits. However, despite its promise, it is unclear what aspects dominate the benefits and whether simply scaling to more examples is the most effective way of improving many-shot ICL. In this work, we first provide an analysis of the factors driving many-shot ICL, and we find that 1) many-shot performance can still "},"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":"2502.00330","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-01T06:23:24Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"097a643e145c1f69b739237c7cb7df949c491e53802fe703d4983ad45a3edca6","abstract_canon_sha256":"1cb03b2645d03d1f63816162eb16cad7e2deada270df531443ceaad244e51c0e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:08:32.257518Z","signature_b64":"rSWBIPgbmmhw0eKMEtOORFtJovacmDQq5VZVdLy3OZkSsBJ7QW+z0uWfPPCu87CQsfzFIzVs6FOPTVsz4xMQBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a4d407f45a453cee78e615b2a5c893f8195d4678ccb1710f29b0b0ab9ff52baa","last_reissued_at":"2026-07-05T10:08:32.257105Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:08:32.257105Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Few to Many: Self-Improving Many-Shot Reasoners Through Iterative Optimization and Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Han Zhou, Hootan Nakhost, Ke Jiang, Ruoxi Sun, Sercan \\\"O. Ar{\\i}k, Xingchen Wan","submitted_at":"2025-02-01T06:23:24Z","abstract_excerpt":"Recent advances in long-context large language models (LLMs) have led to the emerging paradigm of many-shot in-context learning (ICL), where it is observed that scaling many more demonstrating examples beyond the conventional few-shot setup in the context can lead to performance benefits. However, despite its promise, it is unclear what aspects dominate the benefits and whether simply scaling to more examples is the most effective way of improving many-shot ICL. In this work, we first provide an analysis of the factors driving many-shot ICL, and we find that 1) many-shot performance can still "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.00330","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/2502.00330/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":"2502.00330","created_at":"2026-07-05T10:08:32.257160+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.00330v1","created_at":"2026-07-05T10:08:32.257160+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.00330","created_at":"2026-07-05T10:08:32.257160+00:00"},{"alias_kind":"pith_short_12","alias_value":"UTKAP5C2IU6O","created_at":"2026-07-05T10:08:32.257160+00:00"},{"alias_kind":"pith_short_16","alias_value":"UTKAP5C2IU6O46HG","created_at":"2026-07-05T10:08:32.257160+00:00"},{"alias_kind":"pith_short_8","alias_value":"UTKAP5C2","created_at":"2026-07-05T10:08:32.257160+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.06974","citing_title":"MILES: Modular Instruction Memory with Learnable Selection for Self-Improving LLM Reasoning","ref_index":37,"is_internal_anchor":true},{"citing_arxiv_id":"2504.02181","citing_title":"A Survey of Scaling in Large Language Model Reasoning","ref_index":201,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A","json":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A.json","graph_json":"https://pith.science/api/pith-number/UTKAP5C2IU6O46HGCWZKLSET7A/graph.json","events_json":"https://pith.science/api/pith-number/UTKAP5C2IU6O46HGCWZKLSET7A/events.json","paper":"https://pith.science/paper/UTKAP5C2"},"agent_actions":{"view_html":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A","download_json":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A.json","view_paper":"https://pith.science/paper/UTKAP5C2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.00330&json=true","fetch_graph":"https://pith.science/api/pith-number/UTKAP5C2IU6O46HGCWZKLSET7A/graph.json","fetch_events":"https://pith.science/api/pith-number/UTKAP5C2IU6O46HGCWZKLSET7A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A/action/storage_attestation","attest_author":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A/action/author_attestation","sign_citation":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A/action/citation_signature","submit_replication":"https://pith.science/pith/UTKAP5C2IU6O46HGCWZKLSET7A/action/replication_record"}},"created_at":"2026-07-05T10:08:32.257160+00:00","updated_at":"2026-07-05T10:08:32.257160+00:00"}