{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:HNN3OEEYADEPUWM663H64FCTE3","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":"7a680a54cd4ca320e76b5d59d3a63440abbefaa17dde1a62a6661b4156248be3","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-22T11:19:14Z","title_canon_sha256":"d6ad46685abef5958f311df6bee41f32a92dc728a847712d22351ce96bcce3ea"},"schema_version":"1.0","source":{"id":"2407.15549","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.15549","created_at":"2026-07-05T11:44:40Z"},{"alias_kind":"arxiv_version","alias_value":"2407.15549v3","created_at":"2026-07-05T11:44:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.15549","created_at":"2026-07-05T11:44:40Z"},{"alias_kind":"pith_short_12","alias_value":"HNN3OEEYADEP","created_at":"2026-07-05T11:44:40Z"},{"alias_kind":"pith_short_16","alias_value":"HNN3OEEYADEPUWM6","created_at":"2026-07-05T11:44:40Z"},{"alias_kind":"pith_short_8","alias_value":"HNN3OEEY","created_at":"2026-07-05T11:44:40Z"}],"graph_snapshots":[{"event_id":"sha256:c8bbad39582f96c9f7a7581e5fe1d126f279a21c098fa8aeae1d9198f392cbac","target":"graph","created_at":"2026-07-05T11:44:40Z","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/2407.15549/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) can often be made to behave in undesirable ways that they are explicitly fine-tuned not to. For example, the LLM red-teaming literature has produced a wide variety of 'jailbreaking' techniques to elicit harmful text from models that were fine-tuned to be harmless. Recent work on red-teaming, model editing, and interpretability suggests that this challenge stems from how (adversarial) fine-tuning largely serves to suppress rather than remove undesirable capabilities from LLMs. Prior work has introduced latent adversarial training (LAT) as a way to improve robustness","authors_text":"Abhay Sheshadri, Aengus Lynch, Aidan Ewart, Asa Cooper Stickland, Cindy Wu, Dylan Hadfield-Menell, Ethan Perez, Henry Sleight, Phillip Guo, Stephen Casper, Vivek Hebbar","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-22T11:19:14Z","title":"Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.15549","kind":"arxiv","version":3},"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":"sha256:4b8f373ff95623d88e2e71e377b448ae1647ea7d1d98d6fe14b2381830cd53ab","target":"record","created_at":"2026-07-05T11:44:40Z","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":"7a680a54cd4ca320e76b5d59d3a63440abbefaa17dde1a62a6661b4156248be3","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-22T11:19:14Z","title_canon_sha256":"d6ad46685abef5958f311df6bee41f32a92dc728a847712d22351ce96bcce3ea"},"schema_version":"1.0","source":{"id":"2407.15549","kind":"arxiv","version":3}},"canonical_sha256":"3b5bb7109800c8fa599ef6cfee145326ec3e153b71d76e7569b688bbfec5e1ab","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3b5bb7109800c8fa599ef6cfee145326ec3e153b71d76e7569b688bbfec5e1ab","first_computed_at":"2026-07-05T11:44:40.263371Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:44:40.263371Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"gwKy3GA7Rqj8BBO+TnUgmHVKtPi7/n7oSh8+z/tNA/j1qdtMn603LWAOE+HpXi7/Q31tLIdBTRkT0ptDo1okAw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:44:40.263880Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.15549","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4b8f373ff95623d88e2e71e377b448ae1647ea7d1d98d6fe14b2381830cd53ab","sha256:c8bbad39582f96c9f7a7581e5fe1d126f279a21c098fa8aeae1d9198f392cbac"],"state_sha256":"34e9c5bfdd7e373113657446fb177e2e97bdadc479faeb404f5334765714bdb4"}