{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ZDY5Z3HVFGDYUL3QITYDEAEC7E","short_pith_number":"pith:ZDY5Z3HV","schema_version":"1.0","canonical_sha256":"c8f1dcecf529878a2f7044f0320082f9330ae4ef20c546a9d20e5ba5c0639370","source":{"kind":"arxiv","id":"2405.02764","version":2},"attestation_state":"computed","paper":{"title":"Assessing Adversarial Robustness of Large Language Models: An Empirical Study","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Roger Wattenhofer, Xiaochen Zheng, Zeyu Yang, Zhao Meng","submitted_at":"2024-05-04T22:00:28Z","abstract_excerpt":"Large Language Models (LLMs) have revolutionized natural language processing, but their robustness against adversarial attacks remains a critical concern. We presents a novel white-box style attack approach that exposes vulnerabilities in leading open-source LLMs, including Llama, OPT, and T5. We assess the impact of model size, structure, and fine-tuning strategies on their resistance to adversarial perturbations. Our comprehensive evaluation across five diverse text classification tasks establishes a new benchmark for LLM robustness. The findings of this study have far-reaching implications "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2405.02764","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-04T22:00:28Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"cf6c81fc8fde220e53054e00c744aa01dd0ddcd3bc4ffbd791115e781509ab81","abstract_canon_sha256":"025f9da112c73f90c4f83270af0a1ab41c8bd3838b9796bbe9af8d9ed18db983"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:06:26.152028Z","signature_b64":"Ql9ROOmEeQxh/1ypIORGBGZZAyWpoYP9SjqWN5lf/MI7aZ1QkdtGWY+i2MxoQl2chuYq2GVKtBvIBg/9R2+3Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c8f1dcecf529878a2f7044f0320082f9330ae4ef20c546a9d20e5ba5c0639370","last_reissued_at":"2026-07-05T09:06:26.151490Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:06:26.151490Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Assessing Adversarial Robustness of Large Language Models: An Empirical Study","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Roger Wattenhofer, Xiaochen Zheng, Zeyu Yang, Zhao Meng","submitted_at":"2024-05-04T22:00:28Z","abstract_excerpt":"Large Language Models (LLMs) have revolutionized natural language processing, but their robustness against adversarial attacks remains a critical concern. We presents a novel white-box style attack approach that exposes vulnerabilities in leading open-source LLMs, including Llama, OPT, and T5. We assess the impact of model size, structure, and fine-tuning strategies on their resistance to adversarial perturbations. Our comprehensive evaluation across five diverse text classification tasks establishes a new benchmark for LLM robustness. The findings of this study have far-reaching implications "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.02764","kind":"arxiv","version":2},"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/2405.02764/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2405.02764","created_at":"2026-07-05T09:06:26.151555+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.02764v2","created_at":"2026-07-05T09:06:26.151555+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.02764","created_at":"2026-07-05T09:06:26.151555+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZDY5Z3HVFGDY","created_at":"2026-07-05T09:06:26.151555+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZDY5Z3HVFGDYUL3Q","created_at":"2026-07-05T09:06:26.151555+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZDY5Z3HV","created_at":"2026-07-05T09:06:26.151555+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.01800","citing_title":"Do LLMs Truly Generalize in the Molecular Domain? A Perturbation-Based Analysis","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13439","citing_title":"S-GBT: Smooth Growth Bound Tensor for Certified Robustness Against Word Substitution Attacks in NLP","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04177","citing_title":"Are LLMs Ready for Conflict Monitoring? Empirical Evidence from West Africa","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E","json":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E.json","graph_json":"https://pith.science/api/pith-number/ZDY5Z3HVFGDYUL3QITYDEAEC7E/graph.json","events_json":"https://pith.science/api/pith-number/ZDY5Z3HVFGDYUL3QITYDEAEC7E/events.json","paper":"https://pith.science/paper/ZDY5Z3HV"},"agent_actions":{"view_html":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E","download_json":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E.json","view_paper":"https://pith.science/paper/ZDY5Z3HV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.02764&json=true","fetch_graph":"https://pith.science/api/pith-number/ZDY5Z3HVFGDYUL3QITYDEAEC7E/graph.json","fetch_events":"https://pith.science/api/pith-number/ZDY5Z3HVFGDYUL3QITYDEAEC7E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E/action/storage_attestation","attest_author":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E/action/author_attestation","sign_citation":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E/action/citation_signature","submit_replication":"https://pith.science/pith/ZDY5Z3HVFGDYUL3QITYDEAEC7E/action/replication_record"}},"created_at":"2026-07-05T09:06:26.151555+00:00","updated_at":"2026-07-05T09:06:26.151555+00:00"}