{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:5BJXHONL5AGLDMNFTZMYCNXO4Z","short_pith_number":"pith:5BJXHONL","canonical_record":{"source":{"id":"2605.12964","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T03:58:01Z","cross_cats_sorted":[],"title_canon_sha256":"21c39b6d70b359de02bc8c2ed213c6b0f6bcb042d60390d4e87feb195cdbc4cd","abstract_canon_sha256":"6da0e2f77d32b5f0e27f3f7233909305891727fac98fe6e79b9775b9fb466964"},"schema_version":"1.0"},"canonical_sha256":"e85373b9abe80cb1b1a59e598136eee6583fdb2efb700f875ca150f0df8989af","source":{"kind":"arxiv","id":"2605.12964","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12964","created_at":"2026-05-18T03:09:09Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12964v1","created_at":"2026-05-18T03:09:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12964","created_at":"2026-05-18T03:09:09Z"},{"alias_kind":"pith_short_12","alias_value":"5BJXHONL5AGL","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"5BJXHONL5AGLDMNF","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"5BJXHONL","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:5BJXHONL5AGLDMNFTZMYCNXO4Z","target":"record","payload":{"canonical_record":{"source":{"id":"2605.12964","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T03:58:01Z","cross_cats_sorted":[],"title_canon_sha256":"21c39b6d70b359de02bc8c2ed213c6b0f6bcb042d60390d4e87feb195cdbc4cd","abstract_canon_sha256":"6da0e2f77d32b5f0e27f3f7233909305891727fac98fe6e79b9775b9fb466964"},"schema_version":"1.0"},"canonical_sha256":"e85373b9abe80cb1b1a59e598136eee6583fdb2efb700f875ca150f0df8989af","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:09:09.100345Z","signature_b64":"76c9CUgfP6hiwrJCq7OPh5m6D5ygM89BAHdLPqF+vBSPLbCGEEW3s4+1p+2YO3PngUPZMFk9F8aBYb/FojsACg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e85373b9abe80cb1b1a59e598136eee6583fdb2efb700f875ca150f0df8989af","last_reissued_at":"2026-05-18T03:09:09.099499Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:09:09.099499Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.12964","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":"2026-05-18T03:09:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2Y+o74NsJPU2Kzr9VLYTHKTP8eao31wK7RFswnrzoGnZXmWyQl2Sn8Vo5At2+lrqnRSlimRLw5X5UdLriNyzAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-31T18:27:18.150785Z"},"content_sha256":"c90fe35ba8a2261030f9e14ef6bdc1cf5165fe69d1b4a47ed98f95c96982d7e4","schema_version":"1.0","event_id":"sha256:c90fe35ba8a2261030f9e14ef6bdc1cf5165fe69d1b4a47ed98f95c96982d7e4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:5BJXHONL5AGLDMNFTZMYCNXO4Z","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Asymmetric Flow Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"AsymFlow achieves 1.57 FID on ImageNet by predicting noise only in a low-rank subspace while recovering full-dimensional velocity analytically.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Gordon Wetzstein, Hansheng Chen, Jan Ackermann, Leonidas Guibas, Minseo Kim","submitted_at":"2026-05-13T03:58:01Z","abstract_excerpt":"Flow-based generation in high-dimensional spaces is difficult because velocity prediction requires modeling high-dimensional noise, even when data has strong low-rank structure. We present Asymmetric Flow Modeling (AsymFlow), a rank-asymmetric velocity parameterization that restricts noise prediction to a low-rank subspace while keeping data prediction full-dimensional. From this asymmetric prediction, AsymFlow analytically recovers the full-dimensional velocity without changing the network architecture or training/sampling procedures. On ImageNet 256$\\times$256, AsymFlow achieves a leading 1."},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"On ImageNet 256×256, AsymFlow achieves a leading 1.57 FID, outperforming prior DiT/JiT-like pixel diffusion models by a large margin. The pixel AsymFlow model finetuned from FLUX.2 klein 9B establishes a new state of the art for pixel-space text-to-image generation, beating its latent base on HPSv3, DPG-Bench, and GenEval.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The data possesses strong low-rank structure that allows restricting noise prediction to a low-rank subspace without losing critical information needed for accurate full-dimensional velocity recovery.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Asymmetric Flow Modeling restricts noise prediction to a low-rank subspace for high-dimensional flow generation, reaching 1.57 FID on ImageNet 256x256 and new state-of-the-art pixel text-to-image performance via finetuning from latent models.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"AsymFlow achieves 1.57 FID on ImageNet by predicting noise only in a low-rank subspace while recovering full-dimensional velocity analytically.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9aa66385cce8326c021cf83c9e2b2a3879579324447ca4a0c1006662b6a2355c"},"source":{"id":"2605.12964","kind":"arxiv","version":1},"verdict":{"id":"418ec465-cbf9-4351-ab4b-b6a7df13cde7","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:21:49.954041Z","strongest_claim":"On ImageNet 256×256, AsymFlow achieves a leading 1.57 FID, outperforming prior DiT/JiT-like pixel diffusion models by a large margin. The pixel AsymFlow model finetuned from FLUX.2 klein 9B establishes a new state of the art for pixel-space text-to-image generation, beating its latent base on HPSv3, DPG-Bench, and GenEval.","one_line_summary":"Asymmetric Flow Modeling restricts noise prediction to a low-rank subspace for high-dimensional flow generation, reaching 1.57 FID on ImageNet 256x256 and new state-of-the-art pixel text-to-image performance via finetuning from latent models.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The data possesses strong low-rank structure that allows restricting noise prediction to a low-rank subspace without losing critical information needed for accurate full-dimensional velocity recovery.","pith_extraction_headline":"AsymFlow achieves 1.57 FID on ImageNet by predicting noise only in a low-rank subspace while recovering full-dimensional velocity analytically."},"references":{"count":77,"sample":[{"doi":"","year":2023,"title":"Building normalizing flows with stochastic interpolants","work_id":"461610bc-3153-4c5e-89f4-6f2e11fa0112","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"arXiv preprint arXiv:2602.11401 (2026)","work_id":"f794fa7a-d96f-4973-8022-3524cacfd457","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","ref_index":3,"cited_arxiv_id":"2502.13923","is_internal_anchor":true},{"doi":"","year":2023,"title":"All are worth words: A ViT backbone for diffusion models","work_id":"86bac5b7-abcb-4347-99b8-62c082eb2cbf","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Flux.https://github.com/black-forest-labs/flux","work_id":"ff476bd7-afa4-451d-b45d-54207aeb1545","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":77,"snapshot_sha256":"9a428bd07c84078adec7d68a6383c8386fa8e0ed448f83c6642e49e428692ce4","internal_anchors":12},"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":"418ec465-cbf9-4351-ab4b-b6a7df13cde7"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T03:09:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tvZ6t46+P5JCS4b1FeoaGViJIaW5NJgPQNoiQ3bF9qcgB+NbJSQnHkL97giloznsBZTf6EBTi42wDNL0JDYYAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-31T18:27:18.151332Z"},"content_sha256":"bbd1c59aff21bc328fff5698736f7354d0d8163bb5fd7de4774db903dbc7dcc0","schema_version":"1.0","event_id":"sha256:bbd1c59aff21bc328fff5698736f7354d0d8163bb5fd7de4774db903dbc7dcc0"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5BJXHONL5AGLDMNFTZMYCNXO4Z/bundle.json","state_url":"https://pith.science/pith/5BJXHONL5AGLDMNFTZMYCNXO4Z/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5BJXHONL5AGLDMNFTZMYCNXO4Z/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":"2026-05-31T18:27:18Z","links":{"resolver":"https://pith.science/pith/5BJXHONL5AGLDMNFTZMYCNXO4Z","bundle":"https://pith.science/pith/5BJXHONL5AGLDMNFTZMYCNXO4Z/bundle.json","state":"https://pith.science/pith/5BJXHONL5AGLDMNFTZMYCNXO4Z/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5BJXHONL5AGLDMNFTZMYCNXO4Z/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:5BJXHONL5AGLDMNFTZMYCNXO4Z","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":"6da0e2f77d32b5f0e27f3f7233909305891727fac98fe6e79b9775b9fb466964","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T03:58:01Z","title_canon_sha256":"21c39b6d70b359de02bc8c2ed213c6b0f6bcb042d60390d4e87feb195cdbc4cd"},"schema_version":"1.0","source":{"id":"2605.12964","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12964","created_at":"2026-05-18T03:09:09Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12964v1","created_at":"2026-05-18T03:09:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12964","created_at":"2026-05-18T03:09:09Z"},{"alias_kind":"pith_short_12","alias_value":"5BJXHONL5AGL","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"5BJXHONL5AGLDMNF","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"5BJXHONL","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:bbd1c59aff21bc328fff5698736f7354d0d8163bb5fd7de4774db903dbc7dcc0","target":"graph","created_at":"2026-05-18T03:09:09Z","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":"On ImageNet 256×256, AsymFlow achieves a leading 1.57 FID, outperforming prior DiT/JiT-like pixel diffusion models by a large margin. The pixel AsymFlow model finetuned from FLUX.2 klein 9B establishes a new state of the art for pixel-space text-to-image generation, beating its latent base on HPSv3, DPG-Bench, and GenEval."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The data possesses strong low-rank structure that allows restricting noise prediction to a low-rank subspace without losing critical information needed for accurate full-dimensional velocity recovery."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Asymmetric Flow Modeling restricts noise prediction to a low-rank subspace for high-dimensional flow generation, reaching 1.57 FID on ImageNet 256x256 and new state-of-the-art pixel text-to-image performance via finetuning from latent models."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"AsymFlow achieves 1.57 FID on ImageNet by predicting noise only in a low-rank subspace while recovering full-dimensional velocity analytically."}],"snapshot_sha256":"9aa66385cce8326c021cf83c9e2b2a3879579324447ca4a0c1006662b6a2355c"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Flow-based generation in high-dimensional spaces is difficult because velocity prediction requires modeling high-dimensional noise, even when data has strong low-rank structure. We present Asymmetric Flow Modeling (AsymFlow), a rank-asymmetric velocity parameterization that restricts noise prediction to a low-rank subspace while keeping data prediction full-dimensional. From this asymmetric prediction, AsymFlow analytically recovers the full-dimensional velocity without changing the network architecture or training/sampling procedures. On ImageNet 256$\\times$256, AsymFlow achieves a leading 1.","authors_text":"Gordon Wetzstein, Hansheng Chen, Jan Ackermann, Leonidas Guibas, Minseo Kim","cross_cats":[],"headline":"AsymFlow achieves 1.57 FID on ImageNet by predicting noise only in a low-rank subspace while recovering full-dimensional velocity analytically.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T03:58:01Z","title":"Asymmetric Flow Models"},"references":{"count":77,"internal_anchors":12,"resolved_work":77,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Building normalizing flows with stochastic interpolants","work_id":"461610bc-3153-4c5e-89f4-6f2e11fa0112","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"arXiv preprint arXiv:2602.11401 (2026)","work_id":"f794fa7a-d96f-4973-8022-3524cacfd457","year":2026},{"cited_arxiv_id":"2502.13923","doi":"","is_internal_anchor":true,"ref_index":3,"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"All are worth words: A ViT backbone for diffusion models","work_id":"86bac5b7-abcb-4347-99b8-62c082eb2cbf","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Flux.https://github.com/black-forest-labs/flux","work_id":"ff476bd7-afa4-451d-b45d-54207aeb1545","year":2024}],"snapshot_sha256":"9a428bd07c84078adec7d68a6383c8386fa8e0ed448f83c6642e49e428692ce4"},"source":{"id":"2605.12964","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:21:49.954041Z","id":"418ec465-cbf9-4351-ab4b-b6a7df13cde7","model_set":{"reader":"grok-4.3"},"one_line_summary":"Asymmetric Flow Modeling restricts noise prediction to a low-rank subspace for high-dimensional flow generation, reaching 1.57 FID on ImageNet 256x256 and new state-of-the-art pixel text-to-image performance via finetuning from latent models.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"AsymFlow achieves 1.57 FID on ImageNet by predicting noise only in a low-rank subspace while recovering full-dimensional velocity analytically.","strongest_claim":"On ImageNet 256×256, AsymFlow achieves a leading 1.57 FID, outperforming prior DiT/JiT-like pixel diffusion models by a large margin. The pixel AsymFlow model finetuned from FLUX.2 klein 9B establishes a new state of the art for pixel-space text-to-image generation, beating its latent base on HPSv3, DPG-Bench, and GenEval.","weakest_assumption":"The data possesses strong low-rank structure that allows restricting noise prediction to a low-rank subspace without losing critical information needed for accurate full-dimensional velocity recovery."}},"verdict_id":"418ec465-cbf9-4351-ab4b-b6a7df13cde7"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c90fe35ba8a2261030f9e14ef6bdc1cf5165fe69d1b4a47ed98f95c96982d7e4","target":"record","created_at":"2026-05-18T03:09:09Z","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":"6da0e2f77d32b5f0e27f3f7233909305891727fac98fe6e79b9775b9fb466964","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T03:58:01Z","title_canon_sha256":"21c39b6d70b359de02bc8c2ed213c6b0f6bcb042d60390d4e87feb195cdbc4cd"},"schema_version":"1.0","source":{"id":"2605.12964","kind":"arxiv","version":1}},"canonical_sha256":"e85373b9abe80cb1b1a59e598136eee6583fdb2efb700f875ca150f0df8989af","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e85373b9abe80cb1b1a59e598136eee6583fdb2efb700f875ca150f0df8989af","first_computed_at":"2026-05-18T03:09:09.099499Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T03:09:09.099499Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"76c9CUgfP6hiwrJCq7OPh5m6D5ygM89BAHdLPqF+vBSPLbCGEEW3s4+1p+2YO3PngUPZMFk9F8aBYb/FojsACg==","signature_status":"signed_v1","signed_at":"2026-05-18T03:09:09.100345Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.12964","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c90fe35ba8a2261030f9e14ef6bdc1cf5165fe69d1b4a47ed98f95c96982d7e4","sha256:bbd1c59aff21bc328fff5698736f7354d0d8163bb5fd7de4774db903dbc7dcc0"],"state_sha256":"954852e28aac4fb9fb0162b9e033da041398d0ccb28471d04ee539f325a9b83a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ydGXSOlUwWPjx8U8yudDqH2nOIu/54DJT6NXk5SpRdZ3GNPFpyuhgQzVelj09Pd4TOLUeQGo7Jp2g1+WwmMCAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-31T18:27:18.153676Z","bundle_sha256":"d233e7ca3e29f6f24f5ea2f199915154e8818d1893819c2faa8db29f1f302307"}}