{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:INEYQLSGBA2GKU6GZGI5EPGAM4","short_pith_number":"pith:INEYQLSG","canonical_record":{"source":{"id":"2110.13711","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-10-26T14:00:49Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"c97cb31d2ff198fcec054e24aa3917ec2dc26af86a7be9759ea79b86c801c8df","abstract_canon_sha256":"9253b693cd0e79cdbd4dbff206d368cd7a1c53fdb07ef1c595687c5bd9c7ac92"},"schema_version":"1.0"},"canonical_sha256":"4349882e4608346553c6c991d23cc0670f2c67ac745dcccc2ecb56947d8e654e","source":{"kind":"arxiv","id":"2110.13711","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2110.13711","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"arxiv_version","alias_value":"2110.13711v2","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.13711","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"pith_short_12","alias_value":"INEYQLSGBA2G","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"pith_short_16","alias_value":"INEYQLSGBA2GKU6G","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"pith_short_8","alias_value":"INEYQLSG","created_at":"2026-07-05T04:15:11Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:INEYQLSGBA2GKU6GZGI5EPGAM4","target":"record","payload":{"canonical_record":{"source":{"id":"2110.13711","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-10-26T14:00:49Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"c97cb31d2ff198fcec054e24aa3917ec2dc26af86a7be9759ea79b86c801c8df","abstract_canon_sha256":"9253b693cd0e79cdbd4dbff206d368cd7a1c53fdb07ef1c595687c5bd9c7ac92"},"schema_version":"1.0"},"canonical_sha256":"4349882e4608346553c6c991d23cc0670f2c67ac745dcccc2ecb56947d8e654e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:15:11.344370Z","signature_b64":"p2egYmhicuzOJgNRDaaYfqEqRNZyoKbPdHLrWhetNavoQWEUSi8m27HlF+5fXmpT4JxMLN4iHtjIBXlmEdXnCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4349882e4608346553c6c991d23cc0670f2c67ac745dcccc2ecb56947d8e654e","last_reissued_at":"2026-07-05T04:15:11.343902Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:15:11.343902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2110.13711","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T04:15:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"26X1kGJoGCGyNQIJv5x8ogxZkgCRLoCOcjil8ZHKnYRDlIcNak2YlWNMil77a3RmB05TRrhaZF54udbUzp1+Ag==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T09:17:34.568548Z"},"content_sha256":"c4e6ed4a7ac71ac1fc0aeb48f6930b9d5f3cd0bb6768f84691d93161cd8a6061","schema_version":"1.0","event_id":"sha256:c4e6ed4a7ac71ac1fc0aeb48f6930b9d5f3cd0bb6768f84691d93161cd8a6061"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:INEYQLSGBA2GKU6GZGI5EPGAM4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Hierarchical Transformers Are More Efficient Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Christian Szegedy, Henryk Michalewski, {\\L}ukasz Kaiser, Micha{\\l} Tyrolski, Piotr Nawrot, Szymon Tworkowski, Yuhuai Wu","submitted_at":"2021-10-26T14:00:49Z","abstract_excerpt":"Transformer models yield impressive results on many NLP and sequence modeling tasks. Remarkably, Transformers can handle long sequences which allows them to produce long coherent outputs: full paragraphs produced by GPT-3 or well-structured images produced by DALL-E. These large language models are impressive but also very inefficient and costly, which limits their applications and accessibility. We postulate that having an explicit hierarchical architecture is the key to Transformers that efficiently handle long sequences. To verify this claim, we first study different ways to downsample and "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.13711","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/2110.13711/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":"2026-07-05T04:15:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KbcF0f/ft+PCO2U53ljyDVop5SmPGS6G3o56VyH8hHPIAbULIKOnDMFQbGFEcg2XV1asoC0OwSE7LqNbEsLtCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T09:17:34.568924Z"},"content_sha256":"d3d6f8dddea6e7f49ada7a9a4ba15e57924870cafe7ab67c24e443e7fc7e41a2","schema_version":"1.0","event_id":"sha256:d3d6f8dddea6e7f49ada7a9a4ba15e57924870cafe7ab67c24e443e7fc7e41a2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/INEYQLSGBA2GKU6GZGI5EPGAM4/bundle.json","state_url":"https://pith.science/pith/INEYQLSGBA2GKU6GZGI5EPGAM4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/INEYQLSGBA2GKU6GZGI5EPGAM4/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":"2026-07-07T09:17:34Z","links":{"resolver":"https://pith.science/pith/INEYQLSGBA2GKU6GZGI5EPGAM4","bundle":"https://pith.science/pith/INEYQLSGBA2GKU6GZGI5EPGAM4/bundle.json","state":"https://pith.science/pith/INEYQLSGBA2GKU6GZGI5EPGAM4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/INEYQLSGBA2GKU6GZGI5EPGAM4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:INEYQLSGBA2GKU6GZGI5EPGAM4","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":"9253b693cd0e79cdbd4dbff206d368cd7a1c53fdb07ef1c595687c5bd9c7ac92","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-10-26T14:00:49Z","title_canon_sha256":"c97cb31d2ff198fcec054e24aa3917ec2dc26af86a7be9759ea79b86c801c8df"},"schema_version":"1.0","source":{"id":"2110.13711","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2110.13711","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"arxiv_version","alias_value":"2110.13711v2","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.13711","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"pith_short_12","alias_value":"INEYQLSGBA2G","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"pith_short_16","alias_value":"INEYQLSGBA2GKU6G","created_at":"2026-07-05T04:15:11Z"},{"alias_kind":"pith_short_8","alias_value":"INEYQLSG","created_at":"2026-07-05T04:15:11Z"}],"graph_snapshots":[{"event_id":"sha256:d3d6f8dddea6e7f49ada7a9a4ba15e57924870cafe7ab67c24e443e7fc7e41a2","target":"graph","created_at":"2026-07-05T04:15:11Z","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/2110.13711/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Transformer models yield impressive results on many NLP and sequence modeling tasks. Remarkably, Transformers can handle long sequences which allows them to produce long coherent outputs: full paragraphs produced by GPT-3 or well-structured images produced by DALL-E. These large language models are impressive but also very inefficient and costly, which limits their applications and accessibility. We postulate that having an explicit hierarchical architecture is the key to Transformers that efficiently handle long sequences. To verify this claim, we first study different ways to downsample and ","authors_text":"Christian Szegedy, Henryk Michalewski, {\\L}ukasz Kaiser, Micha{\\l} Tyrolski, Piotr Nawrot, Szymon Tworkowski, Yuhuai Wu","cross_cats":["cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-10-26T14:00:49Z","title":"Hierarchical Transformers Are More Efficient Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.13711","kind":"arxiv","version":2},"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:c4e6ed4a7ac71ac1fc0aeb48f6930b9d5f3cd0bb6768f84691d93161cd8a6061","target":"record","created_at":"2026-07-05T04:15:11Z","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":"9253b693cd0e79cdbd4dbff206d368cd7a1c53fdb07ef1c595687c5bd9c7ac92","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-10-26T14:00:49Z","title_canon_sha256":"c97cb31d2ff198fcec054e24aa3917ec2dc26af86a7be9759ea79b86c801c8df"},"schema_version":"1.0","source":{"id":"2110.13711","kind":"arxiv","version":2}},"canonical_sha256":"4349882e4608346553c6c991d23cc0670f2c67ac745dcccc2ecb56947d8e654e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4349882e4608346553c6c991d23cc0670f2c67ac745dcccc2ecb56947d8e654e","first_computed_at":"2026-07-05T04:15:11.343902Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T04:15:11.343902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"p2egYmhicuzOJgNRDaaYfqEqRNZyoKbPdHLrWhetNavoQWEUSi8m27HlF+5fXmpT4JxMLN4iHtjIBXlmEdXnCg==","signature_status":"signed_v1","signed_at":"2026-07-05T04:15:11.344370Z","signed_message":"canonical_sha256_bytes"},"source_id":"2110.13711","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c4e6ed4a7ac71ac1fc0aeb48f6930b9d5f3cd0bb6768f84691d93161cd8a6061","sha256:d3d6f8dddea6e7f49ada7a9a4ba15e57924870cafe7ab67c24e443e7fc7e41a2"],"state_sha256":"c70134422ff7d5c804307701270840d06f465c5394fa3f551ad8cb2e13a7603d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lGfx1MmUE3YwhufBb7HT0kNK3gSDbtRYYdWunHwH1m8US3aXSrDKRY56+UjA/IaTAwrOIHPbQPdpnU28+kZbDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T09:17:34.570844Z","bundle_sha256":"89b946bc09599474ecccb287050a2679f21938a90ac8a931cdf835de4013ac52"}}