{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:VJGZWCI22U3L5ZKBWACAUBFO27","short_pith_number":"pith:VJGZWCI2","schema_version":"1.0","canonical_sha256":"aa4d9b091ad536bee541b0040a04aed7d9ffcef8075c515342173204b7cef98c","source":{"kind":"arxiv","id":"2605.17930","version":1},"attestation_state":"computed","paper":{"title":"InfoFlow: A Framework for Multi-Layer Transformer Analysis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Haotian Jiang, Penghao Yu, Qianxiao Li, Zeyu Bao","submitted_at":"2026-05-18T06:38:39Z","abstract_excerpt":"While the approximation properties of single-layer Transformer architectures have been studied in recent works, a rigorous theoretical understanding of the multi-layer setting remains limited. In this work, we establish that multi-layer Transformers possess fundamentally different approximation capabilities from single-layer ones: for certain retrieval tasks, any single-layer Transformer requires least $\\Omega (\\varepsilon^{-k})$ parameters to achieve precision $\\varepsilon$, where $k$ grows linearly with sequence length $T$, whereas a two-layer Transformer with a single head per layer achieve"},"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":"2605.17930","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-18T06:38:39Z","cross_cats_sorted":[],"title_canon_sha256":"3185af456c8b549b46f3b0a49b9fe5271a623b32fdf7f81326932e00622577f9","abstract_canon_sha256":"a4733d02ccd10c0d4728d6f6f3f59e6148b1286e4cb1c9e6b3dcf310025e3929"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:05:06.632233Z","signature_b64":"lMU8ZvNV1NpOJAz3sE/xUlaSqaCrMn3TJ73VNCoORXzv1YZgh5dOc7Fhw0w0frW42kzozXcEfI/964yo1LHBCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aa4d9b091ad536bee541b0040a04aed7d9ffcef8075c515342173204b7cef98c","last_reissued_at":"2026-05-20T00:05:06.631361Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:05:06.631361Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"InfoFlow: A Framework for Multi-Layer Transformer Analysis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Haotian Jiang, Penghao Yu, Qianxiao Li, Zeyu Bao","submitted_at":"2026-05-18T06:38:39Z","abstract_excerpt":"While the approximation properties of single-layer Transformer architectures have been studied in recent works, a rigorous theoretical understanding of the multi-layer setting remains limited. In this work, we establish that multi-layer Transformers possess fundamentally different approximation capabilities from single-layer ones: for certain retrieval tasks, any single-layer Transformer requires least $\\Omega (\\varepsilon^{-k})$ parameters to achieve precision $\\varepsilon$, where $k$ grows linearly with sequence length $T$, whereas a two-layer Transformer with a single head per layer achieve"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.17930","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/2605.17930/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":"2605.17930","created_at":"2026-05-20T00:05:06.631518+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.17930v1","created_at":"2026-05-20T00:05:06.631518+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17930","created_at":"2026-05-20T00:05:06.631518+00:00"},{"alias_kind":"pith_short_12","alias_value":"VJGZWCI22U3L","created_at":"2026-05-20T00:05:06.631518+00:00"},{"alias_kind":"pith_short_16","alias_value":"VJGZWCI22U3L5ZKB","created_at":"2026-05-20T00:05:06.631518+00:00"},{"alias_kind":"pith_short_8","alias_value":"VJGZWCI2","created_at":"2026-05-20T00:05:06.631518+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27","json":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27.json","graph_json":"https://pith.science/api/pith-number/VJGZWCI22U3L5ZKBWACAUBFO27/graph.json","events_json":"https://pith.science/api/pith-number/VJGZWCI22U3L5ZKBWACAUBFO27/events.json","paper":"https://pith.science/paper/VJGZWCI2"},"agent_actions":{"view_html":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27","download_json":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27.json","view_paper":"https://pith.science/paper/VJGZWCI2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.17930&json=true","fetch_graph":"https://pith.science/api/pith-number/VJGZWCI22U3L5ZKBWACAUBFO27/graph.json","fetch_events":"https://pith.science/api/pith-number/VJGZWCI22U3L5ZKBWACAUBFO27/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27/action/storage_attestation","attest_author":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27/action/author_attestation","sign_citation":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27/action/citation_signature","submit_replication":"https://pith.science/pith/VJGZWCI22U3L5ZKBWACAUBFO27/action/replication_record"}},"created_at":"2026-05-20T00:05:06.631518+00:00","updated_at":"2026-05-20T00:05:06.631518+00:00"}