{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2015:TQTJPPBT3U5DVCSV7KLEZWAR6I","short_pith_number":"pith:TQTJPPBT","canonical_record":{"source":{"id":"1508.07909","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2015-08-31T16:37:31Z","cross_cats_sorted":[],"title_canon_sha256":"6722be1c762778c2577b341007623b5971b87983909ff213c2a3d9e1558fd668","abstract_canon_sha256":"eb4c8c60b6f1e36b8bdf8179098349e27db1aeb18eb99d3ec08aa6ffe2fb8153"},"schema_version":"1.0"},"canonical_sha256":"9c2697bc33dd3a3a8a55fa964cd811f236b952e599ed006333bb29fc4f96efeb","source":{"kind":"arxiv","id":"1508.07909","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1508.07909","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"arxiv_version","alias_value":"1508.07909v5","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1508.07909","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"pith_short_12","alias_value":"TQTJPPBT3U5D","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"pith_short_16","alias_value":"TQTJPPBT3U5DVCSV","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"pith_short_8","alias_value":"TQTJPPBT","created_at":"2026-07-04T21:04:06Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2015:TQTJPPBT3U5DVCSV7KLEZWAR6I","target":"record","payload":{"canonical_record":{"source":{"id":"1508.07909","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2015-08-31T16:37:31Z","cross_cats_sorted":[],"title_canon_sha256":"6722be1c762778c2577b341007623b5971b87983909ff213c2a3d9e1558fd668","abstract_canon_sha256":"eb4c8c60b6f1e36b8bdf8179098349e27db1aeb18eb99d3ec08aa6ffe2fb8153"},"schema_version":"1.0"},"canonical_sha256":"9c2697bc33dd3a3a8a55fa964cd811f236b952e599ed006333bb29fc4f96efeb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-04T21:04:06.804953Z","signature_b64":"4uR1fO6mooHTBvoeYOlIn0LgEWXlfaiyNtLbIlyqEGBaURYqC/emoz2Tm5t4O2qJPTBckZ5UK6vYOLY9nzmQBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9c2697bc33dd3a3a8a55fa964cd811f236b952e599ed006333bb29fc4f96efeb","last_reissued_at":"2026-07-04T21:04:06.804399Z","signature_status":"signed_v1","first_computed_at":"2026-07-04T21:04:06.804399Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1508.07909","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-04T21:04:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/eO40sbInyUYM/MwslYrO9lDAbkhWd6tXb2dJYF+qHVXMSfu1MbysBVlwjfrlf6EAE20WZPxFyzp5ETQpOIOCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-22T17:28:40.189681Z"},"content_sha256":"c5eb9f958dc1c561bc91ebe397745fb79ce394d636949cf6bc0f7ced86bf82e8","schema_version":"1.0","event_id":"sha256:c5eb9f958dc1c561bc91ebe397745fb79ce394d636949cf6bc0f7ced86bf82e8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2015:TQTJPPBT3U5DVCSV7KLEZWAR6I","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Neural Machine Translation of Rare Words with Subword Units","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Encoding rare words as subword sequences enables open-vocabulary neural machine translation.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexandra Birch, Barry Haddow, Rico Sennrich","submitted_at":"2015-08-31T16:37:31Z","abstract_excerpt":"Neural machine translation (NMT) models typically operate with a fixed vocabulary, but translation is an open-vocabulary problem. Previous work addresses the translation of out-of-vocabulary words by backing off to a dictionary. In this paper, we introduce a simpler and more effective approach, making the NMT model capable of open-vocabulary translation by encoding rare and unknown words as sequences of subword units. This is based on the intuition that various word classes are translatable via smaller units than words, for instance names (via character copying or transliteration), compounds ("},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"subword models improve over a back-off dictionary baseline for the WMT 15 translation tasks English-German and English-Russian by 1.1 and 1.3 BLEU, respectively.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That segmenting words into subword units via byte pair encoding or character n-grams preserves enough linguistic structure for the neural model to learn accurate translations of rare words without introducing excessive ambiguity or length.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Subword segmentation via byte pair encoding enables open-vocabulary neural machine translation and improves BLEU scores by 1.1 on English-German and 1.3 on English-Russian WMT 2015 tasks over dictionary back-off baselines.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Encoding rare words as subword sequences enables open-vocabulary neural machine translation.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"3a3c2212b6942165e0a01c2c054dd3f9c66162a0b51502ccb1b6170689b320f5"},"source":{"id":"1508.07909","kind":"arxiv","version":5},"verdict":{"id":"52bf90a1-1de5-4f32-b2e1-2d05e75fd440","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-12T15:10:16.309155Z","strongest_claim":"subword models improve over a back-off dictionary baseline for the WMT 15 translation tasks English-German and English-Russian by 1.1 and 1.3 BLEU, respectively.","one_line_summary":"Subword segmentation via byte pair encoding enables open-vocabulary neural machine translation and improves BLEU scores by 1.1 on English-German and 1.3 on English-Russian WMT 2015 tasks over dictionary back-off baselines.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That segmenting words into subword units via byte pair encoding or character n-grams preserves enough linguistic structure for the neural model to learn accurate translations of rare words without introducing excessive ambiguity or length.","pith_extraction_headline":"Encoding rare words as subword sequences enables open-vocabulary neural machine translation."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1508.07909/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":36,"sample":[{"doi":"","year":2015,"title":"Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2015. Neural Machine Translation by Jointly Learning to Align and Translate . In Proceedings of the International Conference on Learning Representat","work_id":"261bb695-c634-4ba1-ac3c-f1736351e606","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2000,"title":"Issam Bazzi and James R. Glass. 2000. Modeling out-of-vocabulary words for robust speech recognition . In Sixth International Conference on Spoken Language Processing, ICSLP 2000 / INTERSPEECH 2000 , ","work_id":"d778270c-6abf-4143-8ad1-17cad88fd17a","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2014,"title":"Botha and Phil Blunsom","work_id":"bb5edc56-5db8-4bdd-9ae4-f9ddebc53c9e","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2015,"title":"Rohan Chitnis and John DeNero. 2015. Variable-Length Word Encodings for Neural Translation Models . In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP)","work_id":"303cff5d-f8f9-4da7-acb7-badf252e7c24","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2014,"title":"Kyunghyun Cho, Bart van Merrienboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning Phrase Representations using RNN Encoder--Decoder for Statisti","work_id":"a7414f7a-d45d-4280-8157-5f98002b448b","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":36,"snapshot_sha256":"6aaaedc9d315b057cce843f3c01c34abd488ff3229df9a0056abe725e8543aa0","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":"52bf90a1-1de5-4f32-b2e1-2d05e75fd440"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-04T21:04:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"sf2DOlp4RYXOttF3dK3Q7MYm79w9auyZ95Ea4RSb0z4aHJJcNnEgyB+Hx1hfhcuYdiNwvR9RWqE6ONPF8QJcDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-22T17:28:40.190201Z"},"content_sha256":"2756e330a903c783554a8ba6a63ba9319ffed93488834fc6715c54bb3fb466ce","schema_version":"1.0","event_id":"sha256:2756e330a903c783554a8ba6a63ba9319ffed93488834fc6715c54bb3fb466ce"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/TQTJPPBT3U5DVCSV7KLEZWAR6I/bundle.json","state_url":"https://pith.science/pith/TQTJPPBT3U5DVCSV7KLEZWAR6I/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/TQTJPPBT3U5DVCSV7KLEZWAR6I/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":"2026-07-22T17:28:40Z","links":{"resolver":"https://pith.science/pith/TQTJPPBT3U5DVCSV7KLEZWAR6I","bundle":"https://pith.science/pith/TQTJPPBT3U5DVCSV7KLEZWAR6I/bundle.json","state":"https://pith.science/pith/TQTJPPBT3U5DVCSV7KLEZWAR6I/state.json","well_known_bundle":"https://pith.science/.well-known/pith/TQTJPPBT3U5DVCSV7KLEZWAR6I/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2015:TQTJPPBT3U5DVCSV7KLEZWAR6I","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":"eb4c8c60b6f1e36b8bdf8179098349e27db1aeb18eb99d3ec08aa6ffe2fb8153","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2015-08-31T16:37:31Z","title_canon_sha256":"6722be1c762778c2577b341007623b5971b87983909ff213c2a3d9e1558fd668"},"schema_version":"1.0","source":{"id":"1508.07909","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1508.07909","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"arxiv_version","alias_value":"1508.07909v5","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1508.07909","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"pith_short_12","alias_value":"TQTJPPBT3U5D","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"pith_short_16","alias_value":"TQTJPPBT3U5DVCSV","created_at":"2026-07-04T21:04:06Z"},{"alias_kind":"pith_short_8","alias_value":"TQTJPPBT","created_at":"2026-07-04T21:04:06Z"}],"graph_snapshots":[{"event_id":"sha256:2756e330a903c783554a8ba6a63ba9319ffed93488834fc6715c54bb3fb466ce","target":"graph","created_at":"2026-07-04T21:04:06Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"subword models improve over a back-off dictionary baseline for the WMT 15 translation tasks English-German and English-Russian by 1.1 and 1.3 BLEU, respectively."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That segmenting words into subword units via byte pair encoding or character n-grams preserves enough linguistic structure for the neural model to learn accurate translations of rare words without introducing excessive ambiguity or length."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Subword segmentation via byte pair encoding enables open-vocabulary neural machine translation and improves BLEU scores by 1.1 on English-German and 1.3 on English-Russian WMT 2015 tasks over dictionary back-off baselines."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Encoding rare words as subword sequences enables open-vocabulary neural machine translation."}],"snapshot_sha256":"3a3c2212b6942165e0a01c2c054dd3f9c66162a0b51502ccb1b6170689b320f5"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/1508.07909/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Neural machine translation (NMT) models typically operate with a fixed vocabulary, but translation is an open-vocabulary problem. Previous work addresses the translation of out-of-vocabulary words by backing off to a dictionary. In this paper, we introduce a simpler and more effective approach, making the NMT model capable of open-vocabulary translation by encoding rare and unknown words as sequences of subword units. This is based on the intuition that various word classes are translatable via smaller units than words, for instance names (via character copying or transliteration), compounds (","authors_text":"Alexandra Birch, Barry Haddow, Rico Sennrich","cross_cats":[],"headline":"Encoding rare words as subword sequences enables open-vocabulary neural machine translation.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2015-08-31T16:37:31Z","title":"Neural Machine Translation of Rare Words with Subword Units"},"references":{"count":36,"internal_anchors":0,"resolved_work":36,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2015. Neural Machine Translation by Jointly Learning to Align and Translate . In Proceedings of the International Conference on Learning Representat","work_id":"261bb695-c634-4ba1-ac3c-f1736351e606","year":2015},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Issam Bazzi and James R. Glass. 2000. Modeling out-of-vocabulary words for robust speech recognition . In Sixth International Conference on Spoken Language Processing, ICSLP 2000 / INTERSPEECH 2000 , ","work_id":"d778270c-6abf-4143-8ad1-17cad88fd17a","year":2000},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Botha and Phil Blunsom","work_id":"bb5edc56-5db8-4bdd-9ae4-f9ddebc53c9e","year":2014},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Rohan Chitnis and John DeNero. 2015. Variable-Length Word Encodings for Neural Translation Models . In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP)","work_id":"303cff5d-f8f9-4da7-acb7-badf252e7c24","year":2015},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Kyunghyun Cho, Bart van Merrienboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning Phrase Representations using RNN Encoder--Decoder for Statisti","work_id":"a7414f7a-d45d-4280-8157-5f98002b448b","year":2014}],"snapshot_sha256":"6aaaedc9d315b057cce843f3c01c34abd488ff3229df9a0056abe725e8543aa0"},"source":{"id":"1508.07909","kind":"arxiv","version":5},"verdict":{"created_at":"2026-05-12T15:10:16.309155Z","id":"52bf90a1-1de5-4f32-b2e1-2d05e75fd440","model_set":{"reader":"grok-4.3"},"one_line_summary":"Subword segmentation via byte pair encoding enables open-vocabulary neural machine translation and improves BLEU scores by 1.1 on English-German and 1.3 on English-Russian WMT 2015 tasks over dictionary back-off baselines.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Encoding rare words as subword sequences enables open-vocabulary neural machine translation.","strongest_claim":"subword models improve over a back-off dictionary baseline for the WMT 15 translation tasks English-German and English-Russian by 1.1 and 1.3 BLEU, respectively.","weakest_assumption":"That segmenting words into subword units via byte pair encoding or character n-grams preserves enough linguistic structure for the neural model to learn accurate translations of rare words without introducing excessive ambiguity or length."}},"verdict_id":"52bf90a1-1de5-4f32-b2e1-2d05e75fd440"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c5eb9f958dc1c561bc91ebe397745fb79ce394d636949cf6bc0f7ced86bf82e8","target":"record","created_at":"2026-07-04T21:04:06Z","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":"eb4c8c60b6f1e36b8bdf8179098349e27db1aeb18eb99d3ec08aa6ffe2fb8153","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2015-08-31T16:37:31Z","title_canon_sha256":"6722be1c762778c2577b341007623b5971b87983909ff213c2a3d9e1558fd668"},"schema_version":"1.0","source":{"id":"1508.07909","kind":"arxiv","version":5}},"canonical_sha256":"9c2697bc33dd3a3a8a55fa964cd811f236b952e599ed006333bb29fc4f96efeb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9c2697bc33dd3a3a8a55fa964cd811f236b952e599ed006333bb29fc4f96efeb","first_computed_at":"2026-07-04T21:04:06.804399Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-04T21:04:06.804399Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"4uR1fO6mooHTBvoeYOlIn0LgEWXlfaiyNtLbIlyqEGBaURYqC/emoz2Tm5t4O2qJPTBckZ5UK6vYOLY9nzmQBg==","signature_status":"signed_v1","signed_at":"2026-07-04T21:04:06.804953Z","signed_message":"canonical_sha256_bytes"},"source_id":"1508.07909","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c5eb9f958dc1c561bc91ebe397745fb79ce394d636949cf6bc0f7ced86bf82e8","sha256:2756e330a903c783554a8ba6a63ba9319ffed93488834fc6715c54bb3fb466ce"],"state_sha256":"e1f1b41fb5d3d44d582fd1ffc17a7e400d2f9268700b457d2476ef65cbebe8b1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"INXO6nPEHcLITQ1PUmW/tz2VDrokHGUBe8H3jQBQqgXFRJ0af8PuInLrWZGtgZUUGaubH7bc17CgPMBq4rT0Aw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-22T17:28:40.192722Z","bundle_sha256":"1f31b0a9d7522e4320e1257ab24c72d316eed5760e8d881ed0ece7194067611d"}}