{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:O235XRQOVJOLUNFB5VF7GTVWF2","short_pith_number":"pith:O235XRQO","canonical_record":{"source":{"id":"2406.12091","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-17T21:06:00Z","cross_cats_sorted":["cs.CL","cs.CR"],"title_canon_sha256":"bf0069701a10b4b3816e3a8f6874420f4986e57b2433d1bb3376052281f5d25f","abstract_canon_sha256":"0a25fdce08cdd755506e911df31dd9402f38cfd70b23c0b43568d8bac2a6487e"},"schema_version":"1.0"},"canonical_sha256":"76b7dbc60eaa5cba34a1ed4bf34eb62e8c4b9a6187ffae12798080367a2214fe","source":{"kind":"arxiv","id":"2406.12091","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.12091","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"arxiv_version","alias_value":"2406.12091v4","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.12091","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"pith_short_12","alias_value":"O235XRQOVJOL","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"pith_short_16","alias_value":"O235XRQOVJOLUNFB","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"pith_short_8","alias_value":"O235XRQO","created_at":"2026-07-05T11:18:12Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:O235XRQOVJOLUNFB5VF7GTVWF2","target":"record","payload":{"canonical_record":{"source":{"id":"2406.12091","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-17T21:06:00Z","cross_cats_sorted":["cs.CL","cs.CR"],"title_canon_sha256":"bf0069701a10b4b3816e3a8f6874420f4986e57b2433d1bb3376052281f5d25f","abstract_canon_sha256":"0a25fdce08cdd755506e911df31dd9402f38cfd70b23c0b43568d8bac2a6487e"},"schema_version":"1.0"},"canonical_sha256":"76b7dbc60eaa5cba34a1ed4bf34eb62e8c4b9a6187ffae12798080367a2214fe","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:18:12.746853Z","signature_b64":"SgiUIYz6hRaaPYBZ826c3FfSL8HcyFyF9vhjdC+lLjOtqRgnsDPZ7224CFJlj5saKqiyQ/mApnuTWMlCJiVGAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"76b7dbc60eaa5cba34a1ed4bf34eb62e8c4b9a6187ffae12798080367a2214fe","last_reissued_at":"2026-07-05T11:18:12.746275Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:18:12.746275Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.12091","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:18:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3fyOhvqkhwANnM5/duLP5cg2x1DsSFtF23zZY2D3+yH/5w/XiV7MeaVju5rSWcJE9bI0AmeIu9jmxC2HAC3SCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T23:59:46.141427Z"},"content_sha256":"1516653b9772138888943c3ddf153446fba2dfce0343f809d3416a910f03d429","schema_version":"1.0","event_id":"sha256:1516653b9772138888943c3ddf153446fba2dfce0343f809d3416a910f03d429"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:O235XRQOVJOLUNFB5VF7GTVWF2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Is poisoning a real threat to LLM alignment? Maybe more so than you think","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CR"],"primary_cat":"cs.LG","authors_text":"Furong Huang, Pankayaraj Pathmanathan, Souradip Chakraborty, Xiangyu Liu, Yongyuan Liang","submitted_at":"2024-06-17T21:06:00Z","abstract_excerpt":"Recent advancements in Reinforcement Learning with Human Feedback (RLHF) have significantly impacted the alignment of Large Language Models (LLMs). The sensitivity of reinforcement learning algorithms such as Proximal Policy Optimization (PPO) has led to new line work on Direct Policy Optimization (DPO), which treats RLHF in a supervised learning framework. The increased practical use of these RLHF methods warrants an analysis of their vulnerabilities. In this work, we investigate the vulnerabilities of DPO to poisoning attacks under different scenarios and compare the effectiveness of prefere"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.12091","kind":"arxiv","version":4},"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/2406.12091/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":"2026-07-05T11:18:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Z3G/+U+BmA4iy+KZkslGO60krTVIjlpEXXsCAWhS3rr9wCGYsJKnlYqbCMW1maGeU4VYiFat/3IxnXQAw49HDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T23:59:46.141952Z"},"content_sha256":"63c025afb37c150919253688fa3a3f1c0aa96e470ef3a8b9fb4527eae93776e5","schema_version":"1.0","event_id":"sha256:63c025afb37c150919253688fa3a3f1c0aa96e470ef3a8b9fb4527eae93776e5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/O235XRQOVJOLUNFB5VF7GTVWF2/bundle.json","state_url":"https://pith.science/pith/O235XRQOVJOLUNFB5VF7GTVWF2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/O235XRQOVJOLUNFB5VF7GTVWF2/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":"2026-08-05T23:59:46Z","links":{"resolver":"https://pith.science/pith/O235XRQOVJOLUNFB5VF7GTVWF2","bundle":"https://pith.science/pith/O235XRQOVJOLUNFB5VF7GTVWF2/bundle.json","state":"https://pith.science/pith/O235XRQOVJOLUNFB5VF7GTVWF2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/O235XRQOVJOLUNFB5VF7GTVWF2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:O235XRQOVJOLUNFB5VF7GTVWF2","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":"0a25fdce08cdd755506e911df31dd9402f38cfd70b23c0b43568d8bac2a6487e","cross_cats_sorted":["cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-17T21:06:00Z","title_canon_sha256":"bf0069701a10b4b3816e3a8f6874420f4986e57b2433d1bb3376052281f5d25f"},"schema_version":"1.0","source":{"id":"2406.12091","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.12091","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"arxiv_version","alias_value":"2406.12091v4","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.12091","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"pith_short_12","alias_value":"O235XRQOVJOL","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"pith_short_16","alias_value":"O235XRQOVJOLUNFB","created_at":"2026-07-05T11:18:12Z"},{"alias_kind":"pith_short_8","alias_value":"O235XRQO","created_at":"2026-07-05T11:18:12Z"}],"graph_snapshots":[{"event_id":"sha256:63c025afb37c150919253688fa3a3f1c0aa96e470ef3a8b9fb4527eae93776e5","target":"graph","created_at":"2026-07-05T11:18:12Z","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/2406.12091/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advancements in Reinforcement Learning with Human Feedback (RLHF) have significantly impacted the alignment of Large Language Models (LLMs). The sensitivity of reinforcement learning algorithms such as Proximal Policy Optimization (PPO) has led to new line work on Direct Policy Optimization (DPO), which treats RLHF in a supervised learning framework. The increased practical use of these RLHF methods warrants an analysis of their vulnerabilities. In this work, we investigate the vulnerabilities of DPO to poisoning attacks under different scenarios and compare the effectiveness of prefere","authors_text":"Furong Huang, Pankayaraj Pathmanathan, Souradip Chakraborty, Xiangyu Liu, Yongyuan Liang","cross_cats":["cs.CL","cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-17T21:06:00Z","title":"Is poisoning a real threat to LLM alignment? Maybe more so than you think"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.12091","kind":"arxiv","version":4},"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":"sha256:1516653b9772138888943c3ddf153446fba2dfce0343f809d3416a910f03d429","target":"record","created_at":"2026-07-05T11:18:12Z","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":"0a25fdce08cdd755506e911df31dd9402f38cfd70b23c0b43568d8bac2a6487e","cross_cats_sorted":["cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-17T21:06:00Z","title_canon_sha256":"bf0069701a10b4b3816e3a8f6874420f4986e57b2433d1bb3376052281f5d25f"},"schema_version":"1.0","source":{"id":"2406.12091","kind":"arxiv","version":4}},"canonical_sha256":"76b7dbc60eaa5cba34a1ed4bf34eb62e8c4b9a6187ffae12798080367a2214fe","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"76b7dbc60eaa5cba34a1ed4bf34eb62e8c4b9a6187ffae12798080367a2214fe","first_computed_at":"2026-07-05T11:18:12.746275Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:18:12.746275Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SgiUIYz6hRaaPYBZ826c3FfSL8HcyFyF9vhjdC+lLjOtqRgnsDPZ7224CFJlj5saKqiyQ/mApnuTWMlCJiVGAg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:18:12.746853Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.12091","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1516653b9772138888943c3ddf153446fba2dfce0343f809d3416a910f03d429","sha256:63c025afb37c150919253688fa3a3f1c0aa96e470ef3a8b9fb4527eae93776e5"],"state_sha256":"e10bb1e3f8f4f800c685f1db43154f9d4b8447a389d15c8d0fdd8a29e035d607"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2wUZtMqFEC0t+0M9xBaC5UFwluiFcbiK73ReyG8MrvZOWmzlqLECqPvzVThZIJ3PcjQfuYJrxJ9z6rkrk4vUDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T23:59:46.145889Z","bundle_sha256":"1f196ae10f30494122403044ffa732f3876ceb6cea07fc73bc98dfe1084fdb6c"}}