{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:QBZ4F2VSXZWIDRSHHGDRPP7O3F","short_pith_number":"pith:QBZ4F2VS","canonical_record":{"source":{"id":"2510.09260","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-10T10:59:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"78b9abf569615958dc501469e7a49928d0567ad717556986982d43b1dad6d8b9","abstract_canon_sha256":"50360177a178eaf3256d19b7d2c5fdf7d803e867bb318bcec8eae19059eadce9"},"schema_version":"1.0"},"canonical_sha256":"8073c2eab2be6c81c647398717bfeed95e8b3e54eecae2a03abda06a566d7246","source":{"kind":"arxiv","id":"2510.09260","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.09260","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"arxiv_version","alias_value":"2510.09260v2","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.09260","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"pith_short_12","alias_value":"QBZ4F2VSXZWI","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"pith_short_16","alias_value":"QBZ4F2VSXZWIDRSH","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"pith_short_8","alias_value":"QBZ4F2VS","created_at":"2026-06-02T01:03:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:QBZ4F2VSXZWIDRSHHGDRPP7O3F","target":"record","payload":{"canonical_record":{"source":{"id":"2510.09260","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-10T10:59:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"78b9abf569615958dc501469e7a49928d0567ad717556986982d43b1dad6d8b9","abstract_canon_sha256":"50360177a178eaf3256d19b7d2c5fdf7d803e867bb318bcec8eae19059eadce9"},"schema_version":"1.0"},"canonical_sha256":"8073c2eab2be6c81c647398717bfeed95e8b3e54eecae2a03abda06a566d7246","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:15.238483Z","signature_b64":"aUDRCbgMmkK2Xg7yRVT7MzhfAzmhfNDyacdvkgx2NVbD0QbF1UkkhJBeMDaZHtKJjnZ1LR5hNXx2S2Yw05uwBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8073c2eab2be6c81c647398717bfeed95e8b3e54eecae2a03abda06a566d7246","last_reissued_at":"2026-06-02T01:03:15.237878Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:15.237878Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2510.09260","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T01:03:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GUCHLBua+/xNVdscQeaKtU8FuWTsu9jMHcXKbCR9Bsva18avH54JEyhGYt/dbbkEuBIsAQftXbxQ9l+6vrB7BA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T09:39:10.581120Z"},"content_sha256":"4e2792795f78d09d02d5a293b58f0665bc2d9fb5b69ce744215413f83d7bb558","schema_version":"1.0","event_id":"sha256:4e2792795f78d09d02d5a293b58f0665bc2d9fb5b69ce744215413f83d7bb558"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:QBZ4F2VSXZWIDRSHHGDRPP7O3F","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"GREAT: Generalizable Backdoor Attacks in RLHF via Emotion-Aware Trigger Synthesis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CR","authors_text":"Piyush Pant, Subrat Kishore Dutta, Xiao Zhang, Yuelin Xu","submitted_at":"2025-10-10T10:59:14Z","abstract_excerpt":"Recent work has shown that RLHF is highly susceptible to backdoor attacks. However, existing methods often rely on rare tokens or fixed triggers, limiting their impact in realistic scenarios. In this work, we develop GREAT, a novel framework for crafting natural distributional backdoors in RLHF. Specifically, GREAT targets harmful response generation for a vulnerable user subpopulation featured by semantically violent requests paired with emotionally angry triggers. At the core of our framework is a trigger identification pipeline that operates in the model's latent embedding space, leveraging"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.09260","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2510.09260/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","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":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-02T01:03:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"l+c5Qn6hQPDj/ZsNQpKBaUb9mviuWWQrlVZwIrTxdACUoq47qRVRDtWg0JR5yX2kSiXEn/WucnUipSiwMcaSCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T09:39:10.581778Z"},"content_sha256":"cb2aef9550c9bf4a47d5573c7091c3a91971e68aed45d42f2b11ca39a9e983c6","schema_version":"1.0","event_id":"sha256:cb2aef9550c9bf4a47d5573c7091c3a91971e68aed45d42f2b11ca39a9e983c6"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/QBZ4F2VSXZWIDRSHHGDRPP7O3F/bundle.json","state_url":"https://pith.science/pith/QBZ4F2VSXZWIDRSHHGDRPP7O3F/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/QBZ4F2VSXZWIDRSHHGDRPP7O3F/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":"2026-08-08T09:39:10Z","links":{"resolver":"https://pith.science/pith/QBZ4F2VSXZWIDRSHHGDRPP7O3F","bundle":"https://pith.science/pith/QBZ4F2VSXZWIDRSHHGDRPP7O3F/bundle.json","state":"https://pith.science/pith/QBZ4F2VSXZWIDRSHHGDRPP7O3F/state.json","well_known_bundle":"https://pith.science/.well-known/pith/QBZ4F2VSXZWIDRSHHGDRPP7O3F/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:QBZ4F2VSXZWIDRSHHGDRPP7O3F","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":"50360177a178eaf3256d19b7d2c5fdf7d803e867bb318bcec8eae19059eadce9","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-10T10:59:14Z","title_canon_sha256":"78b9abf569615958dc501469e7a49928d0567ad717556986982d43b1dad6d8b9"},"schema_version":"1.0","source":{"id":"2510.09260","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2510.09260","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"arxiv_version","alias_value":"2510.09260v2","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.09260","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"pith_short_12","alias_value":"QBZ4F2VSXZWI","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"pith_short_16","alias_value":"QBZ4F2VSXZWIDRSH","created_at":"2026-06-02T01:03:15Z"},{"alias_kind":"pith_short_8","alias_value":"QBZ4F2VS","created_at":"2026-06-02T01:03:15Z"}],"graph_snapshots":[{"event_id":"sha256:cb2aef9550c9bf4a47d5573c7091c3a91971e68aed45d42f2b11ca39a9e983c6","target":"graph","created_at":"2026-06-02T01:03:15Z","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/2510.09260/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent work has shown that RLHF is highly susceptible to backdoor attacks. However, existing methods often rely on rare tokens or fixed triggers, limiting their impact in realistic scenarios. In this work, we develop GREAT, a novel framework for crafting natural distributional backdoors in RLHF. Specifically, GREAT targets harmful response generation for a vulnerable user subpopulation featured by semantically violent requests paired with emotionally angry triggers. At the core of our framework is a trigger identification pipeline that operates in the model's latent embedding space, leveraging","authors_text":"Piyush Pant, Subrat Kishore Dutta, Xiao Zhang, Yuelin Xu","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-10T10:59:14Z","title":"GREAT: Generalizable Backdoor Attacks in RLHF via Emotion-Aware Trigger Synthesis"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.09260","kind":"arxiv","version":2},"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":"sha256:4e2792795f78d09d02d5a293b58f0665bc2d9fb5b69ce744215413f83d7bb558","target":"record","created_at":"2026-06-02T01:03:15Z","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":"50360177a178eaf3256d19b7d2c5fdf7d803e867bb318bcec8eae19059eadce9","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-10T10:59:14Z","title_canon_sha256":"78b9abf569615958dc501469e7a49928d0567ad717556986982d43b1dad6d8b9"},"schema_version":"1.0","source":{"id":"2510.09260","kind":"arxiv","version":2}},"canonical_sha256":"8073c2eab2be6c81c647398717bfeed95e8b3e54eecae2a03abda06a566d7246","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8073c2eab2be6c81c647398717bfeed95e8b3e54eecae2a03abda06a566d7246","first_computed_at":"2026-06-02T01:03:15.237878Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T01:03:15.237878Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"aUDRCbgMmkK2Xg7yRVT7MzhfAzmhfNDyacdvkgx2NVbD0QbF1UkkhJBeMDaZHtKJjnZ1LR5hNXx2S2Yw05uwBA==","signature_status":"signed_v1","signed_at":"2026-06-02T01:03:15.238483Z","signed_message":"canonical_sha256_bytes"},"source_id":"2510.09260","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4e2792795f78d09d02d5a293b58f0665bc2d9fb5b69ce744215413f83d7bb558","sha256:cb2aef9550c9bf4a47d5573c7091c3a91971e68aed45d42f2b11ca39a9e983c6"],"state_sha256":"8f8d888032a6f15850d9d78da8a05d8908f85a6e23163ce91251ea02e2365aab"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WUIpi2eXEawYt3IAdM9pug9ADtTydG6QeRA443wZzIkyal57rOJp4iDT0byF1RGFOSphXaBsf/1cWRQ4Rdk2CQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T09:39:10.587401Z","bundle_sha256":"c45639c50a5c481ee45d3f0168db23998b58a91eabbb4563174ea86afcdf5f27"}}