{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2020:YRVJVUDV7YART2E6JJW5RAXKPA","short_pith_number":"pith:YRVJVUDV","canonical_record":{"source":{"id":"2012.11747","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-21T23:30:04Z","cross_cats_sorted":[],"title_canon_sha256":"36763f5c3cd00dbdd5933866b4dcd94c1c5e5e4ae8aa50c0d03c5c424fe2250a","abstract_canon_sha256":"493b12ab30443eb4a301728ace841f6ea1d63cae7461995bcff28f2756930362"},"schema_version":"1.0"},"canonical_sha256":"c46a9ad075fe0119e89e4a6dd882ea7825f2a78dffaa9c430dd437689051d4c2","source":{"kind":"arxiv","id":"2012.11747","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2012.11747","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"arxiv_version","alias_value":"2012.11747v3","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.11747","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"pith_short_12","alias_value":"YRVJVUDV7YAR","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"pith_short_16","alias_value":"YRVJVUDV7YART2E6","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"pith_short_8","alias_value":"YRVJVUDV","created_at":"2026-07-05T03:13:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2020:YRVJVUDV7YART2E6JJW5RAXKPA","target":"record","payload":{"canonical_record":{"source":{"id":"2012.11747","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-21T23:30:04Z","cross_cats_sorted":[],"title_canon_sha256":"36763f5c3cd00dbdd5933866b4dcd94c1c5e5e4ae8aa50c0d03c5c424fe2250a","abstract_canon_sha256":"493b12ab30443eb4a301728ace841f6ea1d63cae7461995bcff28f2756930362"},"schema_version":"1.0"},"canonical_sha256":"c46a9ad075fe0119e89e4a6dd882ea7825f2a78dffaa9c430dd437689051d4c2","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:13:24.301687Z","signature_b64":"PM2CAWPCHKqS+arBouMGXQVzcpogGw0f2Jro97AiNkUjqp+/tklOONbKWtQ4uSEMStmWUXYDjJSUNZwimNqRAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c46a9ad075fe0119e89e4a6dd882ea7825f2a78dffaa9c430dd437689051d4c2","last_reissued_at":"2026-07-05T03:13:24.301345Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:13:24.301345Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2012.11747","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T03:13:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PBSe52KST1YO+r0boCXG1EOmLzexdpSb3lzUP2oWQom0X8HWlFGky28XHg7uSo1ZkYcnUK4EnlgUy/TsK89oBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T13:44:52.023606Z"},"content_sha256":"a9ea26faf8d0013c75beb264fc71a36f9b7cde0dc77ea30f5d6576b54be78274","schema_version":"1.0","event_id":"sha256:a9ea26faf8d0013c75beb264fc71a36f9b7cde0dc77ea30f5d6576b54be78274"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2020:YRVJVUDV7YART2E6JJW5RAXKPA","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RealFormer: Transformer Likes Residual Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anirudh Ravula, Bhargav Kanagal, Joshua Ainslie, Ruining He","submitted_at":"2020-12-21T23:30:04Z","abstract_excerpt":"Transformer is the backbone of modern NLP models. In this paper, we propose RealFormer, a simple and generic technique to create Residual Attention Layer Transformer networks that significantly outperform the canonical Transformer and its variants (BERT, ETC, etc.) on a wide spectrum of tasks including Masked Language Modeling, GLUE, SQuAD, Neural Machine Translation, WikiHop, HotpotQA, Natural Questions, and OpenKP. We also observe empirically that RealFormer stabilizes training and leads to models with sparser attention. Source code and pre-trained checkpoints for RealFormer can be found at "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.11747","kind":"arxiv","version":3},"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/2012.11747/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":"2026-07-05T03:13:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MyAyp6TWXMNPSEKdbIDWnj+hCQjeMOAFJ1yMzU1QWRrOsgPg/cNUEOlEO+I7/3m2PIyLtsEBm1RrOZ+CV6+9BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T13:44:52.023977Z"},"content_sha256":"d75276ee78e821385102fc47b7672aaa621fba18fc1347022809ee9c75661994","schema_version":"1.0","event_id":"sha256:d75276ee78e821385102fc47b7672aaa621fba18fc1347022809ee9c75661994"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/YRVJVUDV7YART2E6JJW5RAXKPA/bundle.json","state_url":"https://pith.science/pith/YRVJVUDV7YART2E6JJW5RAXKPA/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/YRVJVUDV7YART2E6JJW5RAXKPA/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":"2026-07-07T13:44:52Z","links":{"resolver":"https://pith.science/pith/YRVJVUDV7YART2E6JJW5RAXKPA","bundle":"https://pith.science/pith/YRVJVUDV7YART2E6JJW5RAXKPA/bundle.json","state":"https://pith.science/pith/YRVJVUDV7YART2E6JJW5RAXKPA/state.json","well_known_bundle":"https://pith.science/.well-known/pith/YRVJVUDV7YART2E6JJW5RAXKPA/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:YRVJVUDV7YART2E6JJW5RAXKPA","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":"493b12ab30443eb4a301728ace841f6ea1d63cae7461995bcff28f2756930362","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-21T23:30:04Z","title_canon_sha256":"36763f5c3cd00dbdd5933866b4dcd94c1c5e5e4ae8aa50c0d03c5c424fe2250a"},"schema_version":"1.0","source":{"id":"2012.11747","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2012.11747","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"arxiv_version","alias_value":"2012.11747v3","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.11747","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"pith_short_12","alias_value":"YRVJVUDV7YAR","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"pith_short_16","alias_value":"YRVJVUDV7YART2E6","created_at":"2026-07-05T03:13:24Z"},{"alias_kind":"pith_short_8","alias_value":"YRVJVUDV","created_at":"2026-07-05T03:13:24Z"}],"graph_snapshots":[{"event_id":"sha256:d75276ee78e821385102fc47b7672aaa621fba18fc1347022809ee9c75661994","target":"graph","created_at":"2026-07-05T03:13:24Z","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/2012.11747/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Transformer is the backbone of modern NLP models. In this paper, we propose RealFormer, a simple and generic technique to create Residual Attention Layer Transformer networks that significantly outperform the canonical Transformer and its variants (BERT, ETC, etc.) on a wide spectrum of tasks including Masked Language Modeling, GLUE, SQuAD, Neural Machine Translation, WikiHop, HotpotQA, Natural Questions, and OpenKP. We also observe empirically that RealFormer stabilizes training and leads to models with sparser attention. Source code and pre-trained checkpoints for RealFormer can be found at ","authors_text":"Anirudh Ravula, Bhargav Kanagal, Joshua Ainslie, Ruining He","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-21T23:30:04Z","title":"RealFormer: Transformer Likes Residual Attention"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.11747","kind":"arxiv","version":3},"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:a9ea26faf8d0013c75beb264fc71a36f9b7cde0dc77ea30f5d6576b54be78274","target":"record","created_at":"2026-07-05T03:13:24Z","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":"493b12ab30443eb4a301728ace841f6ea1d63cae7461995bcff28f2756930362","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-21T23:30:04Z","title_canon_sha256":"36763f5c3cd00dbdd5933866b4dcd94c1c5e5e4ae8aa50c0d03c5c424fe2250a"},"schema_version":"1.0","source":{"id":"2012.11747","kind":"arxiv","version":3}},"canonical_sha256":"c46a9ad075fe0119e89e4a6dd882ea7825f2a78dffaa9c430dd437689051d4c2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c46a9ad075fe0119e89e4a6dd882ea7825f2a78dffaa9c430dd437689051d4c2","first_computed_at":"2026-07-05T03:13:24.301345Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T03:13:24.301345Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PM2CAWPCHKqS+arBouMGXQVzcpogGw0f2Jro97AiNkUjqp+/tklOONbKWtQ4uSEMStmWUXYDjJSUNZwimNqRAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T03:13:24.301687Z","signed_message":"canonical_sha256_bytes"},"source_id":"2012.11747","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a9ea26faf8d0013c75beb264fc71a36f9b7cde0dc77ea30f5d6576b54be78274","sha256:d75276ee78e821385102fc47b7672aaa621fba18fc1347022809ee9c75661994"],"state_sha256":"fd213cda5b7eda1fc70fe63074d7b5e4a8dcd092c6920280014c423f676c55f6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xuEVl4pwysgymZrWDPpRQF7rxU0gfWzm6w2BBC+7z1DIBRc6UdLIbOg3kzS94qikKD4jy/2hMhETzEs9vg0DAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T13:44:52.025875Z","bundle_sha256":"d90627459dc633f5013cba6cd8c5a6a41611810e0c7dd1654f4a507a044563c2"}}