{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:Z6P4AB2NINVDZVFCI4PUF5TUR6","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":"985f073d8abf855c913b398958fa3f66524fca0925e64b6814a2bdbd852e3781","cross_cats_sorted":["cs.CL","cs.CY","cs.HC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-06-17T17:16:45Z","title_canon_sha256":"f6dfaeee0820661afbe8133347ea04c2870083d757d8bda18b5b2078e6503db7"},"schema_version":"1.0","source":{"id":"2406.11757","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.11757","created_at":"2026-07-05T09:24:36Z"},{"alias_kind":"arxiv_version","alias_value":"2406.11757v4","created_at":"2026-07-05T09:24:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.11757","created_at":"2026-07-05T09:24:36Z"},{"alias_kind":"pith_short_12","alias_value":"Z6P4AB2NINVD","created_at":"2026-07-05T09:24:36Z"},{"alias_kind":"pith_short_16","alias_value":"Z6P4AB2NINVDZVFC","created_at":"2026-07-05T09:24:36Z"},{"alias_kind":"pith_short_8","alias_value":"Z6P4AB2N","created_at":"2026-07-05T09:24:36Z"}],"graph_snapshots":[{"event_id":"sha256:62797f36d49061a59b1a16768f6d08c925edf39f5cf282ae915ab86870e003a0","target":"graph","created_at":"2026-07-05T09:24:36Z","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/2406.11757/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"This research introduces STAR, a sociotechnical framework that improves on current best practices for red teaming safety of large language models. STAR makes two key contributions: it enhances steerability by generating parameterised instructions for human red teamers, leading to improved coverage of the risk surface. Parameterised instructions also provide more detailed insights into model failures at no increased cost. Second, STAR improves signal quality by matching demographics to assess harms for specific groups, resulting in more sensitive annotations. STAR further employs a novel step o","authors_text":"Bernat Guillen Pegueroles, Canfer Akbulut, John Mellor, Kristian Lum, Laura Weidinger, Mark Diaz, Mikel Rodriguez, Nahema Marchal, Ravin Kumar, Stevie Bergman, Verena Rieser, William Isaac","cross_cats":["cs.CL","cs.CY","cs.HC"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-06-17T17:16:45Z","title":"STAR: SocioTechnical Approach to Red Teaming Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.11757","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":"sha256:2658e729f4f8e54b60989fa14a83cc3b90fd61c509d98d9ec66b43554158381a","target":"record","created_at":"2026-07-05T09:24:36Z","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":"985f073d8abf855c913b398958fa3f66524fca0925e64b6814a2bdbd852e3781","cross_cats_sorted":["cs.CL","cs.CY","cs.HC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-06-17T17:16:45Z","title_canon_sha256":"f6dfaeee0820661afbe8133347ea04c2870083d757d8bda18b5b2078e6503db7"},"schema_version":"1.0","source":{"id":"2406.11757","kind":"arxiv","version":4}},"canonical_sha256":"cf9fc0074d436a3cd4a2471f42f6748f8cbb2b06d827a3c175a080ff4c08ed08","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"cf9fc0074d436a3cd4a2471f42f6748f8cbb2b06d827a3c175a080ff4c08ed08","first_computed_at":"2026-07-05T09:24:36.145948Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:24:36.145948Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KFn1g48a+uxJglCnzeSTCs5cUvAmDFtxOIQbYWWwpjUb+7xGzawBC0CJa+2OisKlePQIoXyDnb5aoOmBblRgDg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:24:36.146457Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.11757","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2658e729f4f8e54b60989fa14a83cc3b90fd61c509d98d9ec66b43554158381a","sha256:62797f36d49061a59b1a16768f6d08c925edf39f5cf282ae915ab86870e003a0"],"state_sha256":"c65132dd69e9de4a862b001aec6a84250ab9a34dbc32a0ea374e826e648c22da"}