{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:PRY25LEOQNUKHXN7U73XQ66OFY","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":"ff80b6ca95cb6e54a7cbad33f0283b6b52bebc468abcd37d81930b50386a9243","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-06-12T00:01:32Z","title_canon_sha256":"144f74023595510a4507c3ba953d173b618ba375fb5cb3417cf3aae14689ae53"},"schema_version":"1.0","source":{"id":"2406.07778","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.07778","created_at":"2026-07-05T08:57:57Z"},{"alias_kind":"arxiv_version","alias_value":"2406.07778v2","created_at":"2026-07-05T08:57:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.07778","created_at":"2026-07-05T08:57:57Z"},{"alias_kind":"pith_short_12","alias_value":"PRY25LEOQNUK","created_at":"2026-07-05T08:57:57Z"},{"alias_kind":"pith_short_16","alias_value":"PRY25LEOQNUKHXN7","created_at":"2026-07-05T08:57:57Z"},{"alias_kind":"pith_short_8","alias_value":"PRY25LEO","created_at":"2026-07-05T08:57:57Z"}],"graph_snapshots":[{"event_id":"sha256:9b59ae5e9d3d22f8c17e033c7d75af7c6288d9f12ef9c02f92cc045081598097","target":"graph","created_at":"2026-07-05T08:57:57Z","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/2406.07778/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Backdoor data poisoning, inserted within instruction examples used to fine-tune a foundation Large Language Model (LLM) for downstream tasks (\\textit{e.g.,} sentiment prediction), is a serious security concern due to the evasive nature of such attacks. The poisoning is usually in the form of a (seemingly innocuous) trigger word or phrase inserted into a very small fraction of the fine-tuning samples from a target class. Such backdoor attacks can: alter response sentiment, violate censorship, over-refuse (invoke censorship for legitimate queries), inject false content, or trigger nonsense respo","authors_text":"David J. Miller, George Kesidis, Jayaram Raghuram","cross_cats":["cs.AI","cs.CL","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-06-12T00:01:32Z","title":"A Study of Backdoors in Instruction Fine-tuned Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.07778","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":"sha256:41f769cf757023672ed42d8b2fa22dfd5beb6d85202407c7f8a519b5210ac492","target":"record","created_at":"2026-07-05T08:57:57Z","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":"ff80b6ca95cb6e54a7cbad33f0283b6b52bebc468abcd37d81930b50386a9243","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-06-12T00:01:32Z","title_canon_sha256":"144f74023595510a4507c3ba953d173b618ba375fb5cb3417cf3aae14689ae53"},"schema_version":"1.0","source":{"id":"2406.07778","kind":"arxiv","version":2}},"canonical_sha256":"7c71aeac8e8368a3ddbfa7f7787bce2e07ab9c664d1dafbef358901e2d5f5b1e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7c71aeac8e8368a3ddbfa7f7787bce2e07ab9c664d1dafbef358901e2d5f5b1e","first_computed_at":"2026-07-05T08:57:57.156581Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:57:57.156581Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ZKJNoZnCo5+kgT6d73v5Ef+4ZVdrjfHDXerr/dT1f5r9rYfyUB9JxQhqVDOeEvjk3gQulwi76Ef/NDSlEmYSBw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:57:57.157012Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.07778","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:41f769cf757023672ed42d8b2fa22dfd5beb6d85202407c7f8a519b5210ac492","sha256:9b59ae5e9d3d22f8c17e033c7d75af7c6288d9f12ef9c02f92cc045081598097"],"state_sha256":"f278a4af2c386aef721e7f9dadfb54b4ca8821ac5823223a28e7df1d451a3b85"}