{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:4RLN2RG5WEMP4F5E727M3HJJ3J","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":"f8191bfd0cbba752820303714ce47e1b5a5b29acd4f0a05a20392e03dfbd47ed","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-12T09:36:33Z","title_canon_sha256":"b7a38144a197c386e15e5df4decbf4de6322cc3154daeac4c10ac92d3cc99532"},"schema_version":"1.0","source":{"id":"2407.09121","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.09121","created_at":"2026-07-05T11:07:59Z"},{"alias_kind":"arxiv_version","alias_value":"2407.09121v2","created_at":"2026-07-05T11:07:59Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.09121","created_at":"2026-07-05T11:07:59Z"},{"alias_kind":"pith_short_12","alias_value":"4RLN2RG5WEMP","created_at":"2026-07-05T11:07:59Z"},{"alias_kind":"pith_short_16","alias_value":"4RLN2RG5WEMP4F5E","created_at":"2026-07-05T11:07:59Z"},{"alias_kind":"pith_short_8","alias_value":"4RLN2RG5","created_at":"2026-07-05T11:07:59Z"}],"graph_snapshots":[{"event_id":"sha256:e593de09ebe271681564480a0c557c7ff8ece101475275aa6020faae5b74824b","target":"graph","created_at":"2026-07-05T11:07:59Z","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/2407.09121/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"This study addresses a critical gap in safety tuning practices for Large Language Models (LLMs) by identifying and tackling a refusal position bias within safety tuning data, which compromises the models' ability to appropriately refuse generating unsafe content. We introduce a novel approach, Decoupled Refusal Training (DeRTa), designed to empower LLMs to refuse compliance to harmful prompts at any response position, significantly enhancing their safety capabilities. DeRTa incorporates two novel components: (1) Maximum Likelihood Estimation (MLE) with Harmful Response Prefix, which trains mod","authors_text":"Jen-tse Huang, Jiahao Xu, Pinjia He, Tian Liang, Wenxiang Jiao, Wenxuan Wang, Youliang Yuan, Zhaopeng Tu","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-12T09:36:33Z","title":"Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.09121","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":"sha256:50d090053f38ddf4677820968a0e65b11a0f7d0a9686d758bd4cc4b78bdf3657","target":"record","created_at":"2026-07-05T11:07:59Z","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":"f8191bfd0cbba752820303714ce47e1b5a5b29acd4f0a05a20392e03dfbd47ed","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-07-12T09:36:33Z","title_canon_sha256":"b7a38144a197c386e15e5df4decbf4de6322cc3154daeac4c10ac92d3cc99532"},"schema_version":"1.0","source":{"id":"2407.09121","kind":"arxiv","version":2}},"canonical_sha256":"e456dd44ddb118fe17a4febecd9d29da6bb7c51eab1b629216a2388116d43365","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e456dd44ddb118fe17a4febecd9d29da6bb7c51eab1b629216a2388116d43365","first_computed_at":"2026-07-05T11:07:59.287516Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:07:59.287516Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"RgljBMvfQ/Wgud97Gq6Q/YoO0wqHNcqBIvNKTBFCtv6seotoS1Km6M26jWvjFJNsB3kss7H8OZ+rgiHbcToICg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:07:59.288068Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.09121","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:50d090053f38ddf4677820968a0e65b11a0f7d0a9686d758bd4cc4b78bdf3657","sha256:e593de09ebe271681564480a0c557c7ff8ece101475275aa6020faae5b74824b"],"state_sha256":"df5c779f0ce34299e37ac954d71541074e87c4f520110e5fbe50e70a28fe7024"}