{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:BYO6ZUZAKDHZTZFUY5RFGTNQCS","short_pith_number":"pith:BYO6ZUZA","canonical_record":{"source":{"id":"2605.15220","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T02:29:19Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"47ccfa359dc3ecc490055107686b0da1b7f1ea016203b4f3c159ab97bd172d75","abstract_canon_sha256":"b02916e3588c09fe7233174d7c1f1ad36ab5296ff41f7fdcf464af241eab45ab"},"schema_version":"1.0"},"canonical_sha256":"0e1decd32050cf99e4b4c762534db0148411cc5d3cdbb5e35a9dae5a7b84a2cb","source":{"kind":"arxiv","id":"2605.15220","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15220","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15220v1","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15220","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"pith_short_12","alias_value":"BYO6ZUZAKDHZ","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"pith_short_16","alias_value":"BYO6ZUZAKDHZTZFU","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"pith_short_8","alias_value":"BYO6ZUZA","created_at":"2026-05-20T00:00:46Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:BYO6ZUZAKDHZTZFUY5RFGTNQCS","target":"record","payload":{"canonical_record":{"source":{"id":"2605.15220","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T02:29:19Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"47ccfa359dc3ecc490055107686b0da1b7f1ea016203b4f3c159ab97bd172d75","abstract_canon_sha256":"b02916e3588c09fe7233174d7c1f1ad36ab5296ff41f7fdcf464af241eab45ab"},"schema_version":"1.0"},"canonical_sha256":"0e1decd32050cf99e4b4c762534db0148411cc5d3cdbb5e35a9dae5a7b84a2cb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:00:46.925585Z","signature_b64":"8mRdunXC7RKHGARKLYYYEHbgBWs/+Gf9HyxAeG4CinT9PKAqyooRfhrnlNuGsNpY9H9dm9X55WCiAz9sIXI1BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0e1decd32050cf99e4b4c762534db0148411cc5d3cdbb5e35a9dae5a7b84a2cb","last_reissued_at":"2026-05-20T00:00:46.924763Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:00:46.924763Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.15220","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":"2026-05-20T00:00:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vTyZxF2iLRTZV+DLBQjDyeCMSaksxsEdHoFlgXqFTN8rPL7rN81kTQIXFu+N7fW6fzEUepndKyO67fPtpsFiBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T02:28:30.522979Z"},"content_sha256":"f29ab024e2cafb3f1523227578679aa965a90a306f4f26d85d26c9f6203c82aa","schema_version":"1.0","event_id":"sha256:f29ab024e2cafb3f1523227578679aa965a90a306f4f26d85d26c9f6203c82aa"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:BYO6ZUZAKDHZTZFUY5RFGTNQCS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Always Learning, Always Mixing: Efficient and Simple Data Mixing All The Time","license":"http://creativecommons.org/licenses/by/4.0/","headline":"OP-Mix simulates candidate data mixtures by interpolating low-rank adapters trained on the current model, enabling efficient mixing across all phases of language model training.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Apurva Gandhi, Kyunghyun Cho, Michael Y. Hu, Pratyusha Sharma, Tal Linzen","submitted_at":"2026-05-13T02:29:19Z","abstract_excerpt":"Data mixing decides how to combine different sources or types of data and is a consequential problem throughout language model training. In pretraining, data composition is a key determinant of model quality; in continual learning and adaptation, it governs what is retained and acquired. Yet existing data mixing methods address only one phase of this lifecycle at a time: some require smaller proxy models tied to a single training phase, others assume a fixed domain set, and continual learning lacks principled guidance altogether. We argue that data mixing is fundamentally an online decision ma"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"OP-Mix consistently finds near-optimal mixtures while using a fraction of the compute of the baselines. In pretraining, OP-Mix improves upon training without mixing by 6.3% in average perplexity. For continual learning, OP-Mix matches the performance of both retraining and on-policy distillation while using 66% and 95% less overall compute, respectively.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That interpolating between low-rank adapters trained directly on the current model accurately simulates the effect of different data mixtures on the full model's learning dynamics without requiring separate proxy models or fixed domain assumptions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"OP-Mix is an on-policy data mixing method that uses low-rank adapter interpolation to find near-optimal data mixtures throughout language model training with reduced compute.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"OP-Mix simulates candidate data mixtures by interpolating low-rank adapters trained on the current model, enabling efficient mixing across all phases of language model training.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f0e401400c40593734ab6146fe81262d4461ff44949ce60afd8468106bd5071c"},"source":{"id":"2605.15220","kind":"arxiv","version":1},"verdict":{"id":"339ecbef-5898-4b8e-8945-a92cebd9b566","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T17:02:49.231906Z","strongest_claim":"OP-Mix consistently finds near-optimal mixtures while using a fraction of the compute of the baselines. In pretraining, OP-Mix improves upon training without mixing by 6.3% in average perplexity. For continual learning, OP-Mix matches the performance of both retraining and on-policy distillation while using 66% and 95% less overall compute, respectively.","one_line_summary":"OP-Mix is an on-policy data mixing method that uses low-rank adapter interpolation to find near-optimal data mixtures throughout language model training with reduced compute.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That interpolating between low-rank adapters trained directly on the current model accurately simulates the effect of different data mixtures on the full model's learning dynamics without requiring separate proxy models or fixed domain assumptions.","pith_extraction_headline":"OP-Mix simulates candidate data mixtures by interpolating low-rank adapters trained on the current model, enabling efficient mixing across all phases of language model training."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15220/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-19T20:41:57.269061Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T17:31:18.496309Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T17:16:24.179511Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.836965Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"c0daef4d60fef3c276f6427a85831c7024e6628576d3c550e5b261f82f9c3496"},"references":{"count":61,"sample":[{"doi":"","year":null,"title":"arXiv preprint arXiv:2602.12237 , year=","work_id":"2fcf1d83-43dc-487a-b86b-0be8003af640","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"The Thirteenth International Conference on Learning Representations , year=","work_id":"5111e23f-4be7-4449-999d-41444f6801f8","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2020,"title":"and Carbin, Michael , title =","work_id":"3da34d91-1a11-4e8e-ae4b-f8c2205042e5","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.5281/zenodo.12608602","year":null,"title":"doi:10.5281/zenodo.12608602 , url =","work_id":"2c57f331-c45e-4b28-9887-1931db7da39f","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Proceedings of the 39th International Conference on Machine Learning , pages =","work_id":"22a5d760-a374-4cf9-8fab-d0244b82548c","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":61,"snapshot_sha256":"028979116fbec170898b7ec7eb8b2b6f934a127cf8c13bd65016324b511b4302","internal_anchors":3},"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":"339ecbef-5898-4b8e-8945-a92cebd9b566"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:00:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2RqRX1iYsLLsP8XDPhetxucJQsCo1qKBQNFTWii32hsvbP8aKTQhO/KMLdXgyNFNU5lLkD6zBWeykrKv8TBQAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T02:28:30.523723Z"},"content_sha256":"9ed0dce10327dd5f500acd7ea29e21cc8d56f048d1dca159331b933e565802ec","schema_version":"1.0","event_id":"sha256:9ed0dce10327dd5f500acd7ea29e21cc8d56f048d1dca159331b933e565802ec"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BYO6ZUZAKDHZTZFUY5RFGTNQCS/bundle.json","state_url":"https://pith.science/pith/BYO6ZUZAKDHZTZFUY5RFGTNQCS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BYO6ZUZAKDHZTZFUY5RFGTNQCS/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":"2026-05-26T02:28:30Z","links":{"resolver":"https://pith.science/pith/BYO6ZUZAKDHZTZFUY5RFGTNQCS","bundle":"https://pith.science/pith/BYO6ZUZAKDHZTZFUY5RFGTNQCS/bundle.json","state":"https://pith.science/pith/BYO6ZUZAKDHZTZFUY5RFGTNQCS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BYO6ZUZAKDHZTZFUY5RFGTNQCS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:BYO6ZUZAKDHZTZFUY5RFGTNQCS","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":"b02916e3588c09fe7233174d7c1f1ad36ab5296ff41f7fdcf464af241eab45ab","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T02:29:19Z","title_canon_sha256":"47ccfa359dc3ecc490055107686b0da1b7f1ea016203b4f3c159ab97bd172d75"},"schema_version":"1.0","source":{"id":"2605.15220","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15220","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15220v1","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15220","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"pith_short_12","alias_value":"BYO6ZUZAKDHZ","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"pith_short_16","alias_value":"BYO6ZUZAKDHZTZFU","created_at":"2026-05-20T00:00:46Z"},{"alias_kind":"pith_short_8","alias_value":"BYO6ZUZA","created_at":"2026-05-20T00:00:46Z"}],"graph_snapshots":[{"event_id":"sha256:9ed0dce10327dd5f500acd7ea29e21cc8d56f048d1dca159331b933e565802ec","target":"graph","created_at":"2026-05-20T00:00:46Z","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":"OP-Mix consistently finds near-optimal mixtures while using a fraction of the compute of the baselines. In pretraining, OP-Mix improves upon training without mixing by 6.3% in average perplexity. For continual learning, OP-Mix matches the performance of both retraining and on-policy distillation while using 66% and 95% less overall compute, respectively."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That interpolating between low-rank adapters trained directly on the current model accurately simulates the effect of different data mixtures on the full model's learning dynamics without requiring separate proxy models or fixed domain assumptions."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"OP-Mix is an on-policy data mixing method that uses low-rank adapter interpolation to find near-optimal data mixtures throughout language model training with reduced compute."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"OP-Mix simulates candidate data mixtures by interpolating low-rank adapters trained on the current model, enabling efficient mixing across all phases of language model training."}],"snapshot_sha256":"f0e401400c40593734ab6146fe81262d4461ff44949ce60afd8468106bd5071c"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T20:41:57.269061Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T17:31:18.496309Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T17:16:24.179511Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.836965Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.15220/integrity.json","findings":[],"snapshot_sha256":"c0daef4d60fef3c276f6427a85831c7024e6628576d3c550e5b261f82f9c3496","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Data mixing decides how to combine different sources or types of data and is a consequential problem throughout language model training. In pretraining, data composition is a key determinant of model quality; in continual learning and adaptation, it governs what is retained and acquired. Yet existing data mixing methods address only one phase of this lifecycle at a time: some require smaller proxy models tied to a single training phase, others assume a fixed domain set, and continual learning lacks principled guidance altogether. We argue that data mixing is fundamentally an online decision ma","authors_text":"Apurva Gandhi, Kyunghyun Cho, Michael Y. Hu, Pratyusha Sharma, Tal Linzen","cross_cats":["cs.AI","cs.LG"],"headline":"OP-Mix simulates candidate data mixtures by interpolating low-rank adapters trained on the current model, enabling efficient mixing across all phases of language model training.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T02:29:19Z","title":"Always Learning, Always Mixing: Efficient and Simple Data Mixing All The Time"},"references":{"count":61,"internal_anchors":3,"resolved_work":61,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"arXiv preprint arXiv:2602.12237 , year=","work_id":"2fcf1d83-43dc-487a-b86b-0be8003af640","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"The Thirteenth International Conference on Learning Representations , year=","work_id":"5111e23f-4be7-4449-999d-41444f6801f8","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"and Carbin, Michael , title =","work_id":"3da34d91-1a11-4e8e-ae4b-f8c2205042e5","year":2020},{"cited_arxiv_id":"","doi":"10.5281/zenodo.12608602","is_internal_anchor":false,"ref_index":4,"title":"doi:10.5281/zenodo.12608602 , url =","work_id":"2c57f331-c45e-4b28-9887-1931db7da39f","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Proceedings of the 39th International Conference on Machine Learning , pages =","work_id":"22a5d760-a374-4cf9-8fab-d0244b82548c","year":2022}],"snapshot_sha256":"028979116fbec170898b7ec7eb8b2b6f934a127cf8c13bd65016324b511b4302"},"source":{"id":"2605.15220","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T17:02:49.231906Z","id":"339ecbef-5898-4b8e-8945-a92cebd9b566","model_set":{"reader":"grok-4.3"},"one_line_summary":"OP-Mix is an on-policy data mixing method that uses low-rank adapter interpolation to find near-optimal data mixtures throughout language model training with reduced compute.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"OP-Mix simulates candidate data mixtures by interpolating low-rank adapters trained on the current model, enabling efficient mixing across all phases of language model training.","strongest_claim":"OP-Mix consistently finds near-optimal mixtures while using a fraction of the compute of the baselines. In pretraining, OP-Mix improves upon training without mixing by 6.3% in average perplexity. For continual learning, OP-Mix matches the performance of both retraining and on-policy distillation while using 66% and 95% less overall compute, respectively.","weakest_assumption":"That interpolating between low-rank adapters trained directly on the current model accurately simulates the effect of different data mixtures on the full model's learning dynamics without requiring separate proxy models or fixed domain assumptions."}},"verdict_id":"339ecbef-5898-4b8e-8945-a92cebd9b566"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:f29ab024e2cafb3f1523227578679aa965a90a306f4f26d85d26c9f6203c82aa","target":"record","created_at":"2026-05-20T00:00:46Z","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":"b02916e3588c09fe7233174d7c1f1ad36ab5296ff41f7fdcf464af241eab45ab","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T02:29:19Z","title_canon_sha256":"47ccfa359dc3ecc490055107686b0da1b7f1ea016203b4f3c159ab97bd172d75"},"schema_version":"1.0","source":{"id":"2605.15220","kind":"arxiv","version":1}},"canonical_sha256":"0e1decd32050cf99e4b4c762534db0148411cc5d3cdbb5e35a9dae5a7b84a2cb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0e1decd32050cf99e4b4c762534db0148411cc5d3cdbb5e35a9dae5a7b84a2cb","first_computed_at":"2026-05-20T00:00:46.924763Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:00:46.924763Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"8mRdunXC7RKHGARKLYYYEHbgBWs/+Gf9HyxAeG4CinT9PKAqyooRfhrnlNuGsNpY9H9dm9X55WCiAz9sIXI1BQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:00:46.925585Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.15220","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f29ab024e2cafb3f1523227578679aa965a90a306f4f26d85d26c9f6203c82aa","sha256:9ed0dce10327dd5f500acd7ea29e21cc8d56f048d1dca159331b933e565802ec"],"state_sha256":"5fd726ee9903d9fae3332b81974fe4166918d94b8d82a96c09e7891f701e54ea"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TLig1k1RhGmGSxB+TgMRbOa0JeZdNLZnRdFS5JeNkTHc+TgqJyIRdD+KPYRuQ/qxVrA1Ym+jHmyk21LTLXgQBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-26T02:28:30.526530Z","bundle_sha256":"4ef268690777efea5a10e37a160b744f5c32167e5ea4bb23b64d89bb723cac8d"}}