{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:BUANOLSA6EH3GYGVRHTC3SGL42","short_pith_number":"pith:BUANOLSA","canonical_record":{"source":{"id":"2402.19442","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-29T18:43:52Z","cross_cats_sorted":["cs.AI","math.OC","math.ST","stat.ML","stat.TH"],"title_canon_sha256":"a010bd88fba21394189b6acc172337e666ef53cc3681352b60ea7d8c85f6deda","abstract_canon_sha256":"bb1a2fa98b97cc639d287c7b1d78ececa2c35b8a92f2bce71ab85fc15f2fa454"},"schema_version":"1.0"},"canonical_sha256":"0d00d72e40f10fb360d589e62dc8cbe687711cd699824507b300925a3b2deb18","source":{"kind":"arxiv","id":"2402.19442","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.19442","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"arxiv_version","alias_value":"2402.19442v2","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.19442","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"pith_short_12","alias_value":"BUANOLSA6EH3","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"pith_short_16","alias_value":"BUANOLSA6EH3GYGV","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"pith_short_8","alias_value":"BUANOLSA","created_at":"2026-07-05T08:29:23Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:BUANOLSA6EH3GYGVRHTC3SGL42","target":"record","payload":{"canonical_record":{"source":{"id":"2402.19442","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-29T18:43:52Z","cross_cats_sorted":["cs.AI","math.OC","math.ST","stat.ML","stat.TH"],"title_canon_sha256":"a010bd88fba21394189b6acc172337e666ef53cc3681352b60ea7d8c85f6deda","abstract_canon_sha256":"bb1a2fa98b97cc639d287c7b1d78ececa2c35b8a92f2bce71ab85fc15f2fa454"},"schema_version":"1.0"},"canonical_sha256":"0d00d72e40f10fb360d589e62dc8cbe687711cd699824507b300925a3b2deb18","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:29:23.397155Z","signature_b64":"24y0fflYpwqIDS2H9kK/g+WSW8C9bJ6TnGzcyJvdbQdyMM18is7aohYiKR6shNHf2DOeBSP2PRd7EfZKEMwwAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0d00d72e40f10fb360d589e62dc8cbe687711cd699824507b300925a3b2deb18","last_reissued_at":"2026-07-05T08:29:23.396659Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:29:23.396659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.19442","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:29:23Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+c0UWKZ8qejTiZ4BDPKAmGOIJH/HLM5dppd3YH+W+lF0ulk0gT3OVEVm210RMBiZ0auyFh/IY8vmTTNaL9/FCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T20:59:14.368993Z"},"content_sha256":"92923305049363f303cfd11f8459e85fef3c1bf656ea5e28fffebd916e3b660e","schema_version":"1.0","event_id":"sha256:92923305049363f303cfd11f8459e85fef3c1bf656ea5e28fffebd916e3b660e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:BUANOLSA6EH3GYGVRHTC3SGL42","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Training Dynamics of Multi-Head Softmax Attention for In-Context Learning: Emergence, Convergence, and Optimality","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","math.OC","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Heejune Sheen, Siyu Chen, Tianhao Wang, Zhuoran Yang","submitted_at":"2024-02-29T18:43:52Z","abstract_excerpt":"We study the dynamics of gradient flow for training a multi-head softmax attention model for in-context learning of multi-task linear regression. We establish the global convergence of gradient flow under suitable choices of initialization. In addition, we prove that an interesting \"task allocation\" phenomenon emerges during the gradient flow dynamics, where each attention head focuses on solving a single task of the multi-task model. Specifically, we prove that the gradient flow dynamics can be split into three phases -- a warm-up phase where the loss decreases rather slowly and the attention"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.19442","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/2402.19442/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:29:23Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"p58Ji2zXNTAf4kBxpDbfBPXltO/HUD0bQK/jOwNkiRT4O+ZynrmQhaHturrG6xcBkbzk6gDPcXTM4E3kHP9GAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T20:59:14.370613Z"},"content_sha256":"a25cc133750adaaf8a4a19934f85700c07f6f8154cb7fadcd16550a753a21a6f","schema_version":"1.0","event_id":"sha256:a25cc133750adaaf8a4a19934f85700c07f6f8154cb7fadcd16550a753a21a6f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BUANOLSA6EH3GYGVRHTC3SGL42/bundle.json","state_url":"https://pith.science/pith/BUANOLSA6EH3GYGVRHTC3SGL42/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BUANOLSA6EH3GYGVRHTC3SGL42/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-07T20:59:14Z","links":{"resolver":"https://pith.science/pith/BUANOLSA6EH3GYGVRHTC3SGL42","bundle":"https://pith.science/pith/BUANOLSA6EH3GYGVRHTC3SGL42/bundle.json","state":"https://pith.science/pith/BUANOLSA6EH3GYGVRHTC3SGL42/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BUANOLSA6EH3GYGVRHTC3SGL42/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:BUANOLSA6EH3GYGVRHTC3SGL42","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"bb1a2fa98b97cc639d287c7b1d78ececa2c35b8a92f2bce71ab85fc15f2fa454","cross_cats_sorted":["cs.AI","math.OC","math.ST","stat.ML","stat.TH"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-29T18:43:52Z","title_canon_sha256":"a010bd88fba21394189b6acc172337e666ef53cc3681352b60ea7d8c85f6deda"},"schema_version":"1.0","source":{"id":"2402.19442","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.19442","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"arxiv_version","alias_value":"2402.19442v2","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.19442","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"pith_short_12","alias_value":"BUANOLSA6EH3","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"pith_short_16","alias_value":"BUANOLSA6EH3GYGV","created_at":"2026-07-05T08:29:23Z"},{"alias_kind":"pith_short_8","alias_value":"BUANOLSA","created_at":"2026-07-05T08:29:23Z"}],"graph_snapshots":[{"event_id":"sha256:a25cc133750adaaf8a4a19934f85700c07f6f8154cb7fadcd16550a753a21a6f","target":"graph","created_at":"2026-07-05T08:29:23Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2402.19442/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We study the dynamics of gradient flow for training a multi-head softmax attention model for in-context learning of multi-task linear regression. We establish the global convergence of gradient flow under suitable choices of initialization. In addition, we prove that an interesting \"task allocation\" phenomenon emerges during the gradient flow dynamics, where each attention head focuses on solving a single task of the multi-task model. Specifically, we prove that the gradient flow dynamics can be split into three phases -- a warm-up phase where the loss decreases rather slowly and the attention","authors_text":"Heejune Sheen, Siyu Chen, Tianhao Wang, Zhuoran Yang","cross_cats":["cs.AI","math.OC","math.ST","stat.ML","stat.TH"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-29T18:43:52Z","title":"Training Dynamics of Multi-Head Softmax Attention for In-Context Learning: Emergence, Convergence, and Optimality"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.19442","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:92923305049363f303cfd11f8459e85fef3c1bf656ea5e28fffebd916e3b660e","target":"record","created_at":"2026-07-05T08:29:23Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"bb1a2fa98b97cc639d287c7b1d78ececa2c35b8a92f2bce71ab85fc15f2fa454","cross_cats_sorted":["cs.AI","math.OC","math.ST","stat.ML","stat.TH"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-29T18:43:52Z","title_canon_sha256":"a010bd88fba21394189b6acc172337e666ef53cc3681352b60ea7d8c85f6deda"},"schema_version":"1.0","source":{"id":"2402.19442","kind":"arxiv","version":2}},"canonical_sha256":"0d00d72e40f10fb360d589e62dc8cbe687711cd699824507b300925a3b2deb18","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0d00d72e40f10fb360d589e62dc8cbe687711cd699824507b300925a3b2deb18","first_computed_at":"2026-07-05T08:29:23.396659Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:29:23.396659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"24y0fflYpwqIDS2H9kK/g+WSW8C9bJ6TnGzcyJvdbQdyMM18is7aohYiKR6shNHf2DOeBSP2PRd7EfZKEMwwAg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:29:23.397155Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.19442","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:92923305049363f303cfd11f8459e85fef3c1bf656ea5e28fffebd916e3b660e","sha256:a25cc133750adaaf8a4a19934f85700c07f6f8154cb7fadcd16550a753a21a6f"],"state_sha256":"93f3f90e9a2080578abde624d4f3626436fd2271ed074d6e9b7f4410a2366eda"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wLOQfsM3vK9GDPsSDeIWWU0LScF3mtd5XMu/OBpPtUc0+iheqgCCp3LWwxfpUnT/va9R343RcZhXqcPUhDAWAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T20:59:14.376783Z","bundle_sha256":"a29cd8bab8a3eb720a09f218f85ee90ccd046bed246a8f24ee038e48cb8b1402"}}