{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:HJK7R5VCVBUGNOVJXV25BLGYVW","short_pith_number":"pith:HJK7R5VC","schema_version":"1.0","canonical_sha256":"3a55f8f6a2a86866baa9bd75d0acd8ad909fcabd73a517ffc4b6855a98a49a0a","source":{"kind":"arxiv","id":"2210.06423","version":2},"attestation_state":"computed","paper":{"title":"Foundation Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Alon Benhaim, Barun Patra, Furu Wei, Hongyu Wang, Li Dong, Payal Bajaj, Saksham Singhal, Shaohan Huang, Shuming Ma, Vishrav Chaudhary, Wenhui Wang, Xia Song, Yu Wu, Zhiliang Peng, Zhun Liu","submitted_at":"2022-10-12T17:16:27Z","abstract_excerpt":"A big convergence of model architectures across language, vision, speech, and multimodal is emerging. However, under the same name \"Transformers\", the above areas use different implementations for better performance, e.g., Post-LayerNorm for BERT, and Pre-LayerNorm for GPT and vision Transformers. We call for the development of Foundation Transformer for true general-purpose modeling, which serves as a go-to architecture for various tasks and modalities with guaranteed training stability. In this work, we introduce a Transformer variant, named Magneto, to fulfill the goal. Specifically, we pro"},"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":"2210.06423","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-12T17:16:27Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"e954c48726f9539c93edeadb6a868978846af88ed2bdecda9f200611b878233b","abstract_canon_sha256":"9345c2d1101a80ba696c58b9bc3a52345b1fd7b8fad136ba60840939156b4d7f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:08:18.392124Z","signature_b64":"EeXfYxTGd3KckLS4DlSkCwgNJppikC9CFIQJiZeJm8yIqg85ZOXTrPyFOv+2ccspCItnYjqqQ2oW75ZyvhuBBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3a55f8f6a2a86866baa9bd75d0acd8ad909fcabd73a517ffc4b6855a98a49a0a","last_reissued_at":"2026-07-05T05:08:18.391618Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:08:18.391618Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Foundation Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Alon Benhaim, Barun Patra, Furu Wei, Hongyu Wang, Li Dong, Payal Bajaj, Saksham Singhal, Shaohan Huang, Shuming Ma, Vishrav Chaudhary, Wenhui Wang, Xia Song, Yu Wu, Zhiliang Peng, Zhun Liu","submitted_at":"2022-10-12T17:16:27Z","abstract_excerpt":"A big convergence of model architectures across language, vision, speech, and multimodal is emerging. However, under the same name \"Transformers\", the above areas use different implementations for better performance, e.g., Post-LayerNorm for BERT, and Pre-LayerNorm for GPT and vision Transformers. We call for the development of Foundation Transformer for true general-purpose modeling, which serves as a go-to architecture for various tasks and modalities with guaranteed training stability. In this work, we introduce a Transformer variant, named Magneto, to fulfill the goal. Specifically, we pro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.06423","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/2210.06423/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":"2210.06423","created_at":"2026-07-05T05:08:18.391684+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.06423v2","created_at":"2026-07-05T05:08:18.391684+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.06423","created_at":"2026-07-05T05:08:18.391684+00:00"},{"alias_kind":"pith_short_12","alias_value":"HJK7R5VCVBUG","created_at":"2026-07-05T05:08:18.391684+00:00"},{"alias_kind":"pith_short_16","alias_value":"HJK7R5VCVBUGNOVJ","created_at":"2026-07-05T05:08:18.391684+00:00"},{"alias_kind":"pith_short_8","alias_value":"HJK7R5VC","created_at":"2026-07-05T05:08:18.391684+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2312.16886","citing_title":"MobileVLM : A Fast, Strong and Open Vision Language Assistant for Mobile Devices","ref_index":120,"is_internal_anchor":false},{"citing_arxiv_id":"2302.14045","citing_title":"Language Is Not All You Need: Aligning Perception with Language Models","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2306.14824","citing_title":"Kosmos-2: Grounding Multimodal Large Language Models to the World","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2307.08621","citing_title":"Retentive Network: A Successor to Transformer for Large Language Models","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW","json":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW.json","graph_json":"https://pith.science/api/pith-number/HJK7R5VCVBUGNOVJXV25BLGYVW/graph.json","events_json":"https://pith.science/api/pith-number/HJK7R5VCVBUGNOVJXV25BLGYVW/events.json","paper":"https://pith.science/paper/HJK7R5VC"},"agent_actions":{"view_html":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW","download_json":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW.json","view_paper":"https://pith.science/paper/HJK7R5VC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.06423&json=true","fetch_graph":"https://pith.science/api/pith-number/HJK7R5VCVBUGNOVJXV25BLGYVW/graph.json","fetch_events":"https://pith.science/api/pith-number/HJK7R5VCVBUGNOVJXV25BLGYVW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW/action/storage_attestation","attest_author":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW/action/author_attestation","sign_citation":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW/action/citation_signature","submit_replication":"https://pith.science/pith/HJK7R5VCVBUGNOVJXV25BLGYVW/action/replication_record"}},"created_at":"2026-07-05T05:08:18.391684+00:00","updated_at":"2026-07-05T05:08:18.391684+00:00"}