{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:YJI57KEFZ6XFRDEC444URU57WB","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":"044f2875ca8ea1cb43e992b200591f661fee26dd93b341872685260b59a5dbae","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-03T16:01:04Z","title_canon_sha256":"7e314dbb397e7bc65995591d8ff09844964eb6d253e2f9788eaeffeb9c588083"},"schema_version":"1.0","source":{"id":"2506.03234","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.03234","created_at":"2026-07-05T11:15:48Z"},{"alias_kind":"arxiv_version","alias_value":"2506.03234v1","created_at":"2026-07-05T11:15:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.03234","created_at":"2026-07-05T11:15:48Z"},{"alias_kind":"pith_short_12","alias_value":"YJI57KEFZ6XF","created_at":"2026-07-05T11:15:48Z"},{"alias_kind":"pith_short_16","alias_value":"YJI57KEFZ6XFRDEC","created_at":"2026-07-05T11:15:48Z"},{"alias_kind":"pith_short_8","alias_value":"YJI57KEF","created_at":"2026-07-05T11:15:48Z"}],"graph_snapshots":[{"event_id":"sha256:546d89d06bdffaf66b139762786c27c53a8e6005267ec98a96982fa97c31be69","target":"graph","created_at":"2026-07-05T11:15:48Z","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/2506.03234/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement Learning from Human Feedback (RLHF) is crucial for aligning text-to-image (T2I) models with human preferences. However, RLHF's feedback mechanism also opens new pathways for adversaries. This paper demonstrates the feasibility of hijacking T2I models by poisoning a small fraction of preference data with natural-appearing examples. Specifically, we propose BadReward, a stealthy clean-label poisoning attack targeting the reward model in multi-modal RLHF. BadReward operates by inducing feature collisions between visually contradicted preference data instances, thereby corrupting the","authors_text":"Cong Wang, Hongwei Yao, Kaiwen Duan, Tong Qiao, Yufei Chen, Zhan Qin, Ziyun Li","cross_cats":["cs.AI","cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-03T16:01:04Z","title":"BadReward: Clean-Label Poisoning of Reward Models in Text-to-Image RLHF"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.03234","kind":"arxiv","version":1},"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":"sha256:926aa97478e8576f1f072bf82842bb5a699b566b9c5e6856c01691f30b160a25","target":"record","created_at":"2026-07-05T11:15:48Z","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":"044f2875ca8ea1cb43e992b200591f661fee26dd93b341872685260b59a5dbae","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-03T16:01:04Z","title_canon_sha256":"7e314dbb397e7bc65995591d8ff09844964eb6d253e2f9788eaeffeb9c588083"},"schema_version":"1.0","source":{"id":"2506.03234","kind":"arxiv","version":1}},"canonical_sha256":"c251dfa885cfae588c82e73948d3bfb04badac45e8cad83b7827f659ad4eb2ab","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c251dfa885cfae588c82e73948d3bfb04badac45e8cad83b7827f659ad4eb2ab","first_computed_at":"2026-07-05T11:15:48.214947Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:15:48.214947Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"CNR2NI6KGyI6Pfh8na65QkIIxdMrUEgXYxbU9pL44S3kpnpk4Ou0XTWmToCyYEmdGUhdGczjJkHtmdArMeq9Cw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:15:48.215447Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.03234","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:926aa97478e8576f1f072bf82842bb5a699b566b9c5e6856c01691f30b160a25","sha256:546d89d06bdffaf66b139762786c27c53a8e6005267ec98a96982fa97c31be69"],"state_sha256":"24f70403dc84ed950fc11a91cc01128b7fd3719bec6c2d03c8b16b19b87c4e81"}