{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:F2KJ5NSL7YWGCLWZHOH44FTI4N","short_pith_number":"pith:F2KJ5NSL","schema_version":"1.0","canonical_sha256":"2e949eb64bfe2c612ed93b8fce1668e371b1e4b18a2c23c15e9e583cc2076886","source":{"kind":"arxiv","id":"2408.04532","version":1},"attestation_state":"computed","paper":{"title":"How Transformers Utilize Multi-Head Attention in In-Context Learning? A Case Study on Sparse Linear Regression","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Difan Zou, Lei Zhao, Xingwu Chen","submitted_at":"2024-08-08T15:33:02Z","abstract_excerpt":"Despite the remarkable success of transformer-based models in various real-world tasks, their underlying mechanisms remain poorly understood. Recent studies have suggested that transformers can implement gradient descent as an in-context learner for linear regression problems and have developed various theoretical analyses accordingly. However, these works mostly focus on the expressive power of transformers by designing specific parameter constructions, lacking a comprehensive understanding of their inherent working mechanisms post-training. In this study, we consider a sparse linear regressi"},"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":"2408.04532","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-08T15:33:02Z","cross_cats_sorted":[],"title_canon_sha256":"c5b469a8f0ec97ef6c6acbaec9d0d0015ea20f7e14d978f34803d6c52f03322b","abstract_canon_sha256":"1877453fcaaf42a7bebc4f19dc1e47148e532d14ebc630b9ef507731da91a51c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:53:34.623005Z","signature_b64":"BGX5hH1uDj5QGDT/zBwBZY8xNlN14eCydX+FG1Q8X4leCz1lQqTVFlG5X9ilH0JP3wfjaYwyZLPrDNnL2KMKAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2e949eb64bfe2c612ed93b8fce1668e371b1e4b18a2c23c15e9e583cc2076886","last_reissued_at":"2026-07-05T08:53:34.622663Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:53:34.622663Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Transformers Utilize Multi-Head Attention in In-Context Learning? A Case Study on Sparse Linear Regression","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Difan Zou, Lei Zhao, Xingwu Chen","submitted_at":"2024-08-08T15:33:02Z","abstract_excerpt":"Despite the remarkable success of transformer-based models in various real-world tasks, their underlying mechanisms remain poorly understood. Recent studies have suggested that transformers can implement gradient descent as an in-context learner for linear regression problems and have developed various theoretical analyses accordingly. However, these works mostly focus on the expressive power of transformers by designing specific parameter constructions, lacking a comprehensive understanding of their inherent working mechanisms post-training. In this study, we consider a sparse linear regressi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.04532","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/2408.04532/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":"2408.04532","created_at":"2026-07-05T08:53:34.622719+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.04532v1","created_at":"2026-07-05T08:53:34.622719+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.04532","created_at":"2026-07-05T08:53:34.622719+00:00"},{"alias_kind":"pith_short_12","alias_value":"F2KJ5NSL7YWG","created_at":"2026-07-05T08:53:34.622719+00:00"},{"alias_kind":"pith_short_16","alias_value":"F2KJ5NSL7YWGCLWZ","created_at":"2026-07-05T08:53:34.622719+00:00"},{"alias_kind":"pith_short_8","alias_value":"F2KJ5NSL","created_at":"2026-07-05T08:53:34.622719+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.07571","citing_title":"Towards Theoretical Understanding of Transformer Test-Time Computing: Investigation on In-Context Linear Regression","ref_index":13,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N","json":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N.json","graph_json":"https://pith.science/api/pith-number/F2KJ5NSL7YWGCLWZHOH44FTI4N/graph.json","events_json":"https://pith.science/api/pith-number/F2KJ5NSL7YWGCLWZHOH44FTI4N/events.json","paper":"https://pith.science/paper/F2KJ5NSL"},"agent_actions":{"view_html":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N","download_json":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N.json","view_paper":"https://pith.science/paper/F2KJ5NSL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.04532&json=true","fetch_graph":"https://pith.science/api/pith-number/F2KJ5NSL7YWGCLWZHOH44FTI4N/graph.json","fetch_events":"https://pith.science/api/pith-number/F2KJ5NSL7YWGCLWZHOH44FTI4N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N/action/storage_attestation","attest_author":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N/action/author_attestation","sign_citation":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N/action/citation_signature","submit_replication":"https://pith.science/pith/F2KJ5NSL7YWGCLWZHOH44FTI4N/action/replication_record"}},"created_at":"2026-07-05T08:53:34.622719+00:00","updated_at":"2026-07-05T08:53:34.622719+00:00"}