{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2015:PCFYDLTODMVFSOODEHBMBTOTEX","short_pith_number":"pith:PCFYDLTO","canonical_record":{"source":{"id":"1512.03385","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2015-12-10T19:51:55Z","cross_cats_sorted":[],"title_canon_sha256":"697404a43beb20f7fff7d7cad129017d947ddc3f622768c25861d5bcd9ae9177","abstract_canon_sha256":"1f2b0fe81acdf1248d06a39083d33225c849991d2be531bec10a0aaeb22a755c"},"schema_version":"1.0"},"canonical_sha256":"788b81ae6e1b2a5939c321c2c0cdd325f377d0fb5a3193bb81fa6d3c44768a05","source":{"kind":"arxiv","id":"1512.03385","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1512.03385","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"arxiv_version","alias_value":"1512.03385v1","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1512.03385","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"pith_short_12","alias_value":"PCFYDLTODMVF","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"pith_short_16","alias_value":"PCFYDLTODMVFSOOD","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"pith_short_8","alias_value":"PCFYDLTO","created_at":"2026-07-04T20:43:20Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2015:PCFYDLTODMVFSOODEHBMBTOTEX","target":"record","payload":{"canonical_record":{"source":{"id":"1512.03385","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2015-12-10T19:51:55Z","cross_cats_sorted":[],"title_canon_sha256":"697404a43beb20f7fff7d7cad129017d947ddc3f622768c25861d5bcd9ae9177","abstract_canon_sha256":"1f2b0fe81acdf1248d06a39083d33225c849991d2be531bec10a0aaeb22a755c"},"schema_version":"1.0"},"canonical_sha256":"788b81ae6e1b2a5939c321c2c0cdd325f377d0fb5a3193bb81fa6d3c44768a05","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-04T20:43:20.476278Z","signature_b64":"gVNx4hN0KMsHjs0VJwB7nRsXJgqM8nEu/Ekq3p7cLs4wwPKgM0f/hLR47/UfuzaVDEO+6WFqLksA6cu3OugbCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"788b81ae6e1b2a5939c321c2c0cdd325f377d0fb5a3193bb81fa6d3c44768a05","last_reissued_at":"2026-07-04T20:43:20.475735Z","signature_status":"signed_v1","first_computed_at":"2026-07-04T20:43:20.475735Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1512.03385","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":"2026-07-04T20:43:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SF7LxzQ5z4KdT+5JrrpD7XuPEpXz8hD0zidn8nnQiOuBmq8K3hzrqT0gEC4W8htJhGf6mpRXBYQTds3eF2/4DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-10T09:30:17.855568Z"},"content_sha256":"26edbc470a8581e671bca6fc2c1690d5ad21c85a4ae46cc43f42e8b9f8ad264a","schema_version":"1.0","event_id":"sha256:26edbc470a8581e671bca6fc2c1690d5ad21c85a4ae46cc43f42e8b9f8ad264a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2015:PCFYDLTODMVFSOODEHBMBTOTEX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Deep Residual Learning for Image Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Residual networks reformulate layers to learn differences from inputs via identity shortcuts, making much deeper training feasible and more accurate.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jian Sun, Kaiming He, Shaoqing Ren, Xiangyu Zhang","submitted_at":"2015-12-10T19:51:55Z","abstract_excerpt":"Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG ne"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the residual functions with identity shortcuts are substantially easier to optimize than the original unreferenced mappings, which the paper supports through experiments but does not prove theoretically.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Residual networks reformulate layers to learn residual functions, enabling effective training of up to 152-layer models that achieve 3.57% error on ImageNet and win ILSVRC 2015.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Residual networks reformulate layers to learn differences from inputs via identity shortcuts, making much deeper training feasible and more accurate.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4ec60011fc0848f655f86be88bbf37ee9730aa80baf90b84029d6a2adac9eb75"},"source":{"id":"1512.03385","kind":"arxiv","version":1},"verdict":{"id":"b5b5bde3-3086-49e2-bb48-80c26a728df5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-11T02:57:14.123676Z","strongest_claim":"We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth.","one_line_summary":"Residual networks reformulate layers to learn residual functions, enabling effective training of up to 152-layer models that achieve 3.57% error on ImageNet and win ILSVRC 2015.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the residual functions with identity shortcuts are substantially easier to optimize than the original unreferenced mappings, which the paper supports through experiments but does not prove theoretically.","pith_extraction_headline":"Residual networks reformulate layers to learn differences from inputs via identity shortcuts, making much deeper training feasible and more accurate."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1512.03385/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":52,"sample":[{"doi":"","year":1994,"title":"Y . Bengio, P. Simard, and P. Frasconi. Learning long-term dependen- cies with gradient descent is difficult. IEEE Transactions on Neural Networks, 5(2):157–166, 1994","work_id":"ca85b4a5-7872-4990-849d-ebde6e745221","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1995,"title":"C. M. Bishop. Neural networks for pattern recognition . Oxford university press, 1995","work_id":"ccefb76f-c5b6-49db-800c-d51470b60ef5","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2000,"title":"W. L. Briggs, S. F. McCormick, et al. A Multigrid Tutorial. Siam, 2000","work_id":"9e910073-1ded-41c6-b232-24989857a9c7","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2011,"title":"K. Chatfield, V . Lempitsky, A. Vedaldi, and A. Zisserman. The devil is in the details: an evaluation of recent feature encoding methods. In BMVC, 2011","work_id":"5adc8414-5588-4c6c-913f-d9dad9367418","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2010,"title":"M. Everingham, L. Van Gool, C. K. Williams, J. Winn, and A. Zis- serman. The Pascal Visual Object Classes (VOC) Challenge. IJCV, pages 303–338, 2010","work_id":"c305d929-5eba-44a5-9dd3-a91930dc5833","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":52,"snapshot_sha256":"0a75757926205bcf3ce280378fdb8f65cc601a2d81cb2098ea01ff6c76aed457","internal_anchors":1},"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":"b5b5bde3-3086-49e2-bb48-80c26a728df5"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-04T20:43:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8PWZiDLP7uQYbYniBxbXEQ9zuysI87dEIvBn2AAN4wPS4MjF2rF0/ixzY5x6HkoqaARQ7CIFq7uAOAuEMCURCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-10T09:30:17.856155Z"},"content_sha256":"f9ae46df44a895f5b6adae9e8e3e1a7ed742d26dec2f239838618da517669a5e","schema_version":"1.0","event_id":"sha256:f9ae46df44a895f5b6adae9e8e3e1a7ed742d26dec2f239838618da517669a5e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/PCFYDLTODMVFSOODEHBMBTOTEX/bundle.json","state_url":"https://pith.science/pith/PCFYDLTODMVFSOODEHBMBTOTEX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/PCFYDLTODMVFSOODEHBMBTOTEX/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":"2026-07-10T09:30:17Z","links":{"resolver":"https://pith.science/pith/PCFYDLTODMVFSOODEHBMBTOTEX","bundle":"https://pith.science/pith/PCFYDLTODMVFSOODEHBMBTOTEX/bundle.json","state":"https://pith.science/pith/PCFYDLTODMVFSOODEHBMBTOTEX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/PCFYDLTODMVFSOODEHBMBTOTEX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2015:PCFYDLTODMVFSOODEHBMBTOTEX","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":"1f2b0fe81acdf1248d06a39083d33225c849991d2be531bec10a0aaeb22a755c","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2015-12-10T19:51:55Z","title_canon_sha256":"697404a43beb20f7fff7d7cad129017d947ddc3f622768c25861d5bcd9ae9177"},"schema_version":"1.0","source":{"id":"1512.03385","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1512.03385","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"arxiv_version","alias_value":"1512.03385v1","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1512.03385","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"pith_short_12","alias_value":"PCFYDLTODMVF","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"pith_short_16","alias_value":"PCFYDLTODMVFSOOD","created_at":"2026-07-04T20:43:20Z"},{"alias_kind":"pith_short_8","alias_value":"PCFYDLTO","created_at":"2026-07-04T20:43:20Z"}],"graph_snapshots":[{"event_id":"sha256:f9ae46df44a895f5b6adae9e8e3e1a7ed742d26dec2f239838618da517669a5e","target":"graph","created_at":"2026-07-04T20:43:20Z","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":"We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the residual functions with identity shortcuts are substantially easier to optimize than the original unreferenced mappings, which the paper supports through experiments but does not prove theoretically."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Residual networks reformulate layers to learn residual functions, enabling effective training of up to 152-layer models that achieve 3.57% error on ImageNet and win ILSVRC 2015."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Residual networks reformulate layers to learn differences from inputs via identity shortcuts, making much deeper training feasible and more accurate."}],"snapshot_sha256":"4ec60011fc0848f655f86be88bbf37ee9730aa80baf90b84029d6a2adac9eb75"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/1512.03385/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG ne","authors_text":"Jian Sun, Kaiming He, Shaoqing Ren, Xiangyu Zhang","cross_cats":[],"headline":"Residual networks reformulate layers to learn differences from inputs via identity shortcuts, making much deeper training feasible and more accurate.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2015-12-10T19:51:55Z","title":"Deep Residual Learning for Image Recognition"},"references":{"count":52,"internal_anchors":1,"resolved_work":52,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Y . Bengio, P. Simard, and P. Frasconi. Learning long-term dependen- cies with gradient descent is difficult. IEEE Transactions on Neural Networks, 5(2):157–166, 1994","work_id":"ca85b4a5-7872-4990-849d-ebde6e745221","year":1994},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"C. M. Bishop. Neural networks for pattern recognition . Oxford university press, 1995","work_id":"ccefb76f-c5b6-49db-800c-d51470b60ef5","year":1995},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"W. L. Briggs, S. F. McCormick, et al. A Multigrid Tutorial. Siam, 2000","work_id":"9e910073-1ded-41c6-b232-24989857a9c7","year":2000},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"K. Chatfield, V . Lempitsky, A. Vedaldi, and A. Zisserman. The devil is in the details: an evaluation of recent feature encoding methods. In BMVC, 2011","work_id":"5adc8414-5588-4c6c-913f-d9dad9367418","year":2011},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"M. Everingham, L. Van Gool, C. K. Williams, J. Winn, and A. Zis- serman. The Pascal Visual Object Classes (VOC) Challenge. IJCV, pages 303–338, 2010","work_id":"c305d929-5eba-44a5-9dd3-a91930dc5833","year":2010}],"snapshot_sha256":"0a75757926205bcf3ce280378fdb8f65cc601a2d81cb2098ea01ff6c76aed457"},"source":{"id":"1512.03385","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-11T02:57:14.123676Z","id":"b5b5bde3-3086-49e2-bb48-80c26a728df5","model_set":{"reader":"grok-4.3"},"one_line_summary":"Residual networks reformulate layers to learn residual functions, enabling effective training of up to 152-layer models that achieve 3.57% error on ImageNet and win ILSVRC 2015.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Residual networks reformulate layers to learn differences from inputs via identity shortcuts, making much deeper training feasible and more accurate.","strongest_claim":"We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth.","weakest_assumption":"That the residual functions with identity shortcuts are substantially easier to optimize than the original unreferenced mappings, which the paper supports through experiments but does not prove theoretically."}},"verdict_id":"b5b5bde3-3086-49e2-bb48-80c26a728df5"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:26edbc470a8581e671bca6fc2c1690d5ad21c85a4ae46cc43f42e8b9f8ad264a","target":"record","created_at":"2026-07-04T20:43:20Z","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":"1f2b0fe81acdf1248d06a39083d33225c849991d2be531bec10a0aaeb22a755c","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2015-12-10T19:51:55Z","title_canon_sha256":"697404a43beb20f7fff7d7cad129017d947ddc3f622768c25861d5bcd9ae9177"},"schema_version":"1.0","source":{"id":"1512.03385","kind":"arxiv","version":1}},"canonical_sha256":"788b81ae6e1b2a5939c321c2c0cdd325f377d0fb5a3193bb81fa6d3c44768a05","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"788b81ae6e1b2a5939c321c2c0cdd325f377d0fb5a3193bb81fa6d3c44768a05","first_computed_at":"2026-07-04T20:43:20.475735Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-04T20:43:20.475735Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"gVNx4hN0KMsHjs0VJwB7nRsXJgqM8nEu/Ekq3p7cLs4wwPKgM0f/hLR47/UfuzaVDEO+6WFqLksA6cu3OugbCg==","signature_status":"signed_v1","signed_at":"2026-07-04T20:43:20.476278Z","signed_message":"canonical_sha256_bytes"},"source_id":"1512.03385","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:26edbc470a8581e671bca6fc2c1690d5ad21c85a4ae46cc43f42e8b9f8ad264a","sha256:f9ae46df44a895f5b6adae9e8e3e1a7ed742d26dec2f239838618da517669a5e"],"state_sha256":"5fce5c55aa2f2ee97fe7b64bf4b5e9538e92b6f3ef39e2890a34137355ce1628"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9b1YiO8+Vzc/wZEWuorocxKTsUlcF5GkmnfXofMrWPBd0eaPgZoRjRmb54nAXGbFnQDrwwZ2zMfKqoefd/fKAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-10T09:30:17.858556Z","bundle_sha256":"4cbf010d9b1ef058a1016cfa37743a33515f70d502c278ca55e9330ed4048395"}}