{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:B5LZDYHPDJMZ7MQP5GM74TLZPH","short_pith_number":"pith:B5LZDYHP","canonical_record":{"source":{"id":"2203.15827","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-29T18:01:24Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e06ab131b2af7dd94616b94599fadba32bc31298e3b1503084333eaacbe93645","abstract_canon_sha256":"017c1cba7b0fcddce92e2aa8454291aa0e3a497229a124b310fd70a07be6c82c"},"schema_version":"1.0"},"canonical_sha256":"0f5791e0ef1a599fb20fe999fe4d7979da7bf53127ef1bcc79939b6287b12b7b","source":{"kind":"arxiv","id":"2203.15827","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2203.15827","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"arxiv_version","alias_value":"2203.15827v1","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.15827","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"pith_short_12","alias_value":"B5LZDYHPDJMZ","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"pith_short_16","alias_value":"B5LZDYHPDJMZ7MQP","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"pith_short_8","alias_value":"B5LZDYHP","created_at":"2026-07-05T04:10:03Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:B5LZDYHPDJMZ7MQP5GM74TLZPH","target":"record","payload":{"canonical_record":{"source":{"id":"2203.15827","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-29T18:01:24Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e06ab131b2af7dd94616b94599fadba32bc31298e3b1503084333eaacbe93645","abstract_canon_sha256":"017c1cba7b0fcddce92e2aa8454291aa0e3a497229a124b310fd70a07be6c82c"},"schema_version":"1.0"},"canonical_sha256":"0f5791e0ef1a599fb20fe999fe4d7979da7bf53127ef1bcc79939b6287b12b7b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:10:03.955660Z","signature_b64":"0PCngBiUkieGoJq4ddP+wgKZRpHqI7mElYoFGTbSFyI69OXst8eQODPTuuXw0fWDDLQGQdD5lfGxPbwxurUXAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0f5791e0ef1a599fb20fe999fe4d7979da7bf53127ef1bcc79939b6287b12b7b","last_reissued_at":"2026-07-05T04:10:03.955158Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:10:03.955158Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2203.15827","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T04:10:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/UwuLcJGQWrGPCOriPf4BhYyUUwRPlHKM8+O2ixBO6G2m5Z6NOWDZmAUB207wcptiSwnBJs4/nGTeNI0CjveDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T09:38:01.114706Z"},"content_sha256":"3b3f25b890850ac0a2eaf327bfb9d6539b068b49e799d98025d51af982eee114","schema_version":"1.0","event_id":"sha256:3b3f25b890850ac0a2eaf327bfb9d6539b068b49e799d98025d51af982eee114"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:B5LZDYHPDJMZ7MQP5GM74TLZPH","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"LinkBERT: Pretraining Language Models with Document Links","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Jure Leskovec, Michihiro Yasunaga, Percy Liang","submitted_at":"2022-03-29T18:01:24Z","abstract_excerpt":"Language model (LM) pretraining can learn various knowledge from text corpora, helping downstream tasks. However, existing methods such as BERT model a single document, and do not capture dependencies or knowledge that span across documents. In this work, we propose LinkBERT, an LM pretraining method that leverages links between documents, e.g., hyperlinks. Given a text corpus, we view it as a graph of documents and create LM inputs by placing linked documents in the same context. We then pretrain the LM with two joint self-supervised objectives: masked language modeling and our new proposal, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.15827","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/2203.15827/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":"2026-07-05T04:10:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3rjNoNefv1sUkOI1ICv968kB2Gwf1yCnEyQbCCTM3VfTNFCx6vhMD+GS+oFlWxa2FH6Zr7sfkPTdhYsY/bmhCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T09:38:01.115202Z"},"content_sha256":"da35d1726de4a94c322dbf04df0fe15f4c100e065147fd062f8dcc61ea188c21","schema_version":"1.0","event_id":"sha256:da35d1726de4a94c322dbf04df0fe15f4c100e065147fd062f8dcc61ea188c21"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/B5LZDYHPDJMZ7MQP5GM74TLZPH/bundle.json","state_url":"https://pith.science/pith/B5LZDYHPDJMZ7MQP5GM74TLZPH/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/B5LZDYHPDJMZ7MQP5GM74TLZPH/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":"2026-08-04T09:38:01Z","links":{"resolver":"https://pith.science/pith/B5LZDYHPDJMZ7MQP5GM74TLZPH","bundle":"https://pith.science/pith/B5LZDYHPDJMZ7MQP5GM74TLZPH/bundle.json","state":"https://pith.science/pith/B5LZDYHPDJMZ7MQP5GM74TLZPH/state.json","well_known_bundle":"https://pith.science/.well-known/pith/B5LZDYHPDJMZ7MQP5GM74TLZPH/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:B5LZDYHPDJMZ7MQP5GM74TLZPH","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":"017c1cba7b0fcddce92e2aa8454291aa0e3a497229a124b310fd70a07be6c82c","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-29T18:01:24Z","title_canon_sha256":"e06ab131b2af7dd94616b94599fadba32bc31298e3b1503084333eaacbe93645"},"schema_version":"1.0","source":{"id":"2203.15827","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2203.15827","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"arxiv_version","alias_value":"2203.15827v1","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.15827","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"pith_short_12","alias_value":"B5LZDYHPDJMZ","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"pith_short_16","alias_value":"B5LZDYHPDJMZ7MQP","created_at":"2026-07-05T04:10:03Z"},{"alias_kind":"pith_short_8","alias_value":"B5LZDYHP","created_at":"2026-07-05T04:10:03Z"}],"graph_snapshots":[{"event_id":"sha256:da35d1726de4a94c322dbf04df0fe15f4c100e065147fd062f8dcc61ea188c21","target":"graph","created_at":"2026-07-05T04:10:03Z","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/2203.15827/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language model (LM) pretraining can learn various knowledge from text corpora, helping downstream tasks. However, existing methods such as BERT model a single document, and do not capture dependencies or knowledge that span across documents. In this work, we propose LinkBERT, an LM pretraining method that leverages links between documents, e.g., hyperlinks. Given a text corpus, we view it as a graph of documents and create LM inputs by placing linked documents in the same context. We then pretrain the LM with two joint self-supervised objectives: masked language modeling and our new proposal, ","authors_text":"Jure Leskovec, Michihiro Yasunaga, Percy Liang","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-29T18:01:24Z","title":"LinkBERT: Pretraining Language Models with Document Links"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.15827","kind":"arxiv","version":1},"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":"sha256:3b3f25b890850ac0a2eaf327bfb9d6539b068b49e799d98025d51af982eee114","target":"record","created_at":"2026-07-05T04:10:03Z","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":"017c1cba7b0fcddce92e2aa8454291aa0e3a497229a124b310fd70a07be6c82c","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-29T18:01:24Z","title_canon_sha256":"e06ab131b2af7dd94616b94599fadba32bc31298e3b1503084333eaacbe93645"},"schema_version":"1.0","source":{"id":"2203.15827","kind":"arxiv","version":1}},"canonical_sha256":"0f5791e0ef1a599fb20fe999fe4d7979da7bf53127ef1bcc79939b6287b12b7b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0f5791e0ef1a599fb20fe999fe4d7979da7bf53127ef1bcc79939b6287b12b7b","first_computed_at":"2026-07-05T04:10:03.955158Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T04:10:03.955158Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"0PCngBiUkieGoJq4ddP+wgKZRpHqI7mElYoFGTbSFyI69OXst8eQODPTuuXw0fWDDLQGQdD5lfGxPbwxurUXAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T04:10:03.955660Z","signed_message":"canonical_sha256_bytes"},"source_id":"2203.15827","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3b3f25b890850ac0a2eaf327bfb9d6539b068b49e799d98025d51af982eee114","sha256:da35d1726de4a94c322dbf04df0fe15f4c100e065147fd062f8dcc61ea188c21"],"state_sha256":"5606c67887d0563426f57990ef87fef663785bc986d83cf663b07e9039be4f09"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"E2iJ9OqCKBeO5Hu0G63JKjqAvIbmh3slaZgeiFQK7R/fWOhE4+zSbX2nB0bzaWpUz6Ws5ksjffpdBGtzGp2yAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T09:38:01.119003Z","bundle_sha256":"cde07fba1a64cfa7555be8d78337331120d3294a1b27b2454857348e21b9124e"}}