{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:CSGQQKG32NFXOPUTI2XQLHW5FL","short_pith_number":"pith:CSGQQKG3","canonical_record":{"source":{"id":"2405.15589","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-24T14:20:09Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"ba0b9bf395da460bdb869dc943d7a7fe8a5a029a92ee806781cbeaf0ca3da718","abstract_canon_sha256":"caec8f9c347aa298f45f4e659879ef107fa55158635765f88cd8e6d5a55e88b0"},"schema_version":"1.0"},"canonical_sha256":"148d0828dbd34b773e9346af059edd2ae8bd9b7a6dc40b3d6cd02e01c0b2b854","source":{"kind":"arxiv","id":"2405.15589","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.15589","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"arxiv_version","alias_value":"2405.15589v3","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.15589","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"pith_short_12","alias_value":"CSGQQKG32NFX","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"pith_short_16","alias_value":"CSGQQKG32NFXOPUT","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"pith_short_8","alias_value":"CSGQQKG3","created_at":"2026-07-05T09:29:28Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:CSGQQKG32NFXOPUTI2XQLHW5FL","target":"record","payload":{"canonical_record":{"source":{"id":"2405.15589","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-24T14:20:09Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"ba0b9bf395da460bdb869dc943d7a7fe8a5a029a92ee806781cbeaf0ca3da718","abstract_canon_sha256":"caec8f9c347aa298f45f4e659879ef107fa55158635765f88cd8e6d5a55e88b0"},"schema_version":"1.0"},"canonical_sha256":"148d0828dbd34b773e9346af059edd2ae8bd9b7a6dc40b3d6cd02e01c0b2b854","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:29:28.987959Z","signature_b64":"3wH4oUIgZPTfW7ZkbZK3+H+3+U/x2FkaX/d4fM89N6B9yi7nsbUefSHLx2O0L/t9QCeeG9iYwEnLkaO8tpEIBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"148d0828dbd34b773e9346af059edd2ae8bd9b7a6dc40b3d6cd02e01c0b2b854","last_reissued_at":"2026-07-05T09:29:28.987479Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:29:28.987479Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.15589","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:29:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"zp6SKDA0UoZdbsj1BxpX4OObiHHRlH5Qd8O0YaW1dnm5QzNfP0qvoUGcxIm5fVdD6T69/0erjabJV/fSEC6ECQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-25T11:32:19.650653Z"},"content_sha256":"cef57c318c5f3116fe31e4f0f116d1619df1e312c5ccf908cbbe261948042109","schema_version":"1.0","event_id":"sha256:cef57c318c5f3116fe31e4f0f116d1619df1e312c5ccf908cbbe261948042109"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:CSGQQKG32NFXOPUTI2XQLHW5FL","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Efficient Adversarial Training in LLMs with Continuous Attacks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"Alessandro Sordoni, Gauthier Gidel, Leo Schwinn, Sophie Xhonneux, Stephan G\\\"unnemann","submitted_at":"2024-05-24T14:20:09Z","abstract_excerpt":"Large language models (LLMs) are vulnerable to adversarial attacks that can bypass their safety guardrails. In many domains, adversarial training has proven to be one of the most promising methods to reliably improve robustness against such attacks. Yet, in the context of LLMs, current methods for adversarial training are hindered by the high computational costs required to perform discrete adversarial attacks at each training iteration. We address this problem by instead calculating adversarial attacks in the continuous embedding space of the LLM, which is orders of magnitudes more efficient."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.15589","kind":"arxiv","version":3},"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/2405.15589/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":"2026-07-05T09:29:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"j3N0VsTUBkAzZNI51dXRzTLXj/ELsgYXsumZeoD34h/iOlqeR51biGsRWznxueRZ9CALp5zLbxuJ6rJRDQRDCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-25T11:32:19.651057Z"},"content_sha256":"0d9374f5e1b78d93f991c80bdf1cb3fccdd0ece48436d1912affcfa077facb92","schema_version":"1.0","event_id":"sha256:0d9374f5e1b78d93f991c80bdf1cb3fccdd0ece48436d1912affcfa077facb92"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CSGQQKG32NFXOPUTI2XQLHW5FL/bundle.json","state_url":"https://pith.science/pith/CSGQQKG32NFXOPUTI2XQLHW5FL/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CSGQQKG32NFXOPUTI2XQLHW5FL/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":"2026-07-25T11:32:19Z","links":{"resolver":"https://pith.science/pith/CSGQQKG32NFXOPUTI2XQLHW5FL","bundle":"https://pith.science/pith/CSGQQKG32NFXOPUTI2XQLHW5FL/bundle.json","state":"https://pith.science/pith/CSGQQKG32NFXOPUTI2XQLHW5FL/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CSGQQKG32NFXOPUTI2XQLHW5FL/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:CSGQQKG32NFXOPUTI2XQLHW5FL","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":"caec8f9c347aa298f45f4e659879ef107fa55158635765f88cd8e6d5a55e88b0","cross_cats_sorted":["cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-24T14:20:09Z","title_canon_sha256":"ba0b9bf395da460bdb869dc943d7a7fe8a5a029a92ee806781cbeaf0ca3da718"},"schema_version":"1.0","source":{"id":"2405.15589","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.15589","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"arxiv_version","alias_value":"2405.15589v3","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.15589","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"pith_short_12","alias_value":"CSGQQKG32NFX","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"pith_short_16","alias_value":"CSGQQKG32NFXOPUT","created_at":"2026-07-05T09:29:28Z"},{"alias_kind":"pith_short_8","alias_value":"CSGQQKG3","created_at":"2026-07-05T09:29:28Z"}],"graph_snapshots":[{"event_id":"sha256:0d9374f5e1b78d93f991c80bdf1cb3fccdd0ece48436d1912affcfa077facb92","target":"graph","created_at":"2026-07-05T09:29:28Z","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/2405.15589/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) are vulnerable to adversarial attacks that can bypass their safety guardrails. In many domains, adversarial training has proven to be one of the most promising methods to reliably improve robustness against such attacks. Yet, in the context of LLMs, current methods for adversarial training are hindered by the high computational costs required to perform discrete adversarial attacks at each training iteration. We address this problem by instead calculating adversarial attacks in the continuous embedding space of the LLM, which is orders of magnitudes more efficient.","authors_text":"Alessandro Sordoni, Gauthier Gidel, Leo Schwinn, Sophie Xhonneux, Stephan G\\\"unnemann","cross_cats":["cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-24T14:20:09Z","title":"Efficient Adversarial Training in LLMs with Continuous Attacks"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.15589","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:cef57c318c5f3116fe31e4f0f116d1619df1e312c5ccf908cbbe261948042109","target":"record","created_at":"2026-07-05T09:29:28Z","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":"caec8f9c347aa298f45f4e659879ef107fa55158635765f88cd8e6d5a55e88b0","cross_cats_sorted":["cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-24T14:20:09Z","title_canon_sha256":"ba0b9bf395da460bdb869dc943d7a7fe8a5a029a92ee806781cbeaf0ca3da718"},"schema_version":"1.0","source":{"id":"2405.15589","kind":"arxiv","version":3}},"canonical_sha256":"148d0828dbd34b773e9346af059edd2ae8bd9b7a6dc40b3d6cd02e01c0b2b854","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"148d0828dbd34b773e9346af059edd2ae8bd9b7a6dc40b3d6cd02e01c0b2b854","first_computed_at":"2026-07-05T09:29:28.987479Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:29:28.987479Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"3wH4oUIgZPTfW7ZkbZK3+H+3+U/x2FkaX/d4fM89N6B9yi7nsbUefSHLx2O0L/t9QCeeG9iYwEnLkaO8tpEIBw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:29:28.987959Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.15589","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:cef57c318c5f3116fe31e4f0f116d1619df1e312c5ccf908cbbe261948042109","sha256:0d9374f5e1b78d93f991c80bdf1cb3fccdd0ece48436d1912affcfa077facb92"],"state_sha256":"163a59a5caf5a073411a8189c990163c1829b9fcb728d9f9c548debfc5121ced"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"e/A7VIyZMw4Vhi/63cyFNMoHc8fmG644xBiTDFa9sirPXoVem2xXx0s4MZE+VR86wf26mIbgXcb30bH0+FxACg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-25T11:32:19.653788Z","bundle_sha256":"9ca45374b06a392316a66e8e5e428cc9f10a735a9f63382da207eb343738eaa3"}}