{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:YNFBNSISMWIYBMDYY54LDWF7QZ","short_pith_number":"pith:YNFBNSIS","schema_version":"1.0","canonical_sha256":"c34a16c912659180b078c778b1d8bf866383c0336c9c7bedb079c69e770d2e20","source":{"kind":"arxiv","id":"2402.09674","version":1},"attestation_state":"computed","paper":{"title":"PAL: Proxy-Guided Black-Box Attack on Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Alexandre Araujo, Chawin Sitawarin, David Wagner, Norman Mu","submitted_at":"2024-02-15T02:54:49Z","abstract_excerpt":"Large Language Models (LLMs) have surged in popularity in recent months, but they have demonstrated concerning capabilities to generate harmful content when manipulated. While techniques like safety fine-tuning aim to minimize harmful use, recent works have shown that LLMs remain vulnerable to attacks that elicit toxic responses. In this work, we introduce the Proxy-Guided Attack on LLMs (PAL), the first optimization-based attack on LLMs in a black-box query-only setting. In particular, it relies on a surrogate model to guide the optimization and a sophisticated loss designed for real-world LL"},"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":"2402.09674","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-15T02:54:49Z","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"title_canon_sha256":"ddd08cbe4da43c72b304c9d8ccdda70931e8e3f90494edc448e125eb831616b9","abstract_canon_sha256":"bbf84ee410771af71fc99d65cfbaea4cf5b1c640eb50bb8428d51a2bd2a929c2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:45:34.738473Z","signature_b64":"yDWFBAFHC5PPmF2TCy/0uJouP0AjZ2+tAWBGYr5Tmzlncprmx7AitAQExta2Z2FTCL7w+iEvXDEGSxVhcexDAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c34a16c912659180b078c778b1d8bf866383c0336c9c7bedb079c69e770d2e20","last_reissued_at":"2026-07-05T07:45:34.738053Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:45:34.738053Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PAL: Proxy-Guided Black-Box Attack on Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Alexandre Araujo, Chawin Sitawarin, David Wagner, Norman Mu","submitted_at":"2024-02-15T02:54:49Z","abstract_excerpt":"Large Language Models (LLMs) have surged in popularity in recent months, but they have demonstrated concerning capabilities to generate harmful content when manipulated. While techniques like safety fine-tuning aim to minimize harmful use, recent works have shown that LLMs remain vulnerable to attacks that elicit toxic responses. In this work, we introduce the Proxy-Guided Attack on LLMs (PAL), the first optimization-based attack on LLMs in a black-box query-only setting. In particular, it relies on a surrogate model to guide the optimization and a sophisticated loss designed for real-world LL"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.09674","kind":"arxiv","version":1},"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/2402.09674/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":"2402.09674","created_at":"2026-07-05T07:45:34.738114+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.09674v1","created_at":"2026-07-05T07:45:34.738114+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.09674","created_at":"2026-07-05T07:45:34.738114+00:00"},{"alias_kind":"pith_short_12","alias_value":"YNFBNSISMWIY","created_at":"2026-07-05T07:45:34.738114+00:00"},{"alias_kind":"pith_short_16","alias_value":"YNFBNSISMWIYBMDY","created_at":"2026-07-05T07:45:34.738114+00:00"},{"alias_kind":"pith_short_8","alias_value":"YNFBNSIS","created_at":"2026-07-05T07:45:34.738114+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08180","citing_title":"Out of Sight: Compression-Aware Content Protection against Agentic Crawlers","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08876","citing_title":"OTora: A Unified Red Teaming Framework for Reasoning-Level Denial-of-Service in LLM Agents","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25073","citing_title":"Security in the Fine-Tuning Lifecycle of Large Language Models: Threats, Defenses,Evaluation, and Future Directions","ref_index":88,"is_internal_anchor":false},{"citing_arxiv_id":"2602.11528","citing_title":"Stop Tracking Me! Proactive Defense Against Attribute Inference Attack in LLMs","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2404.01318","citing_title":"JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2407.04295","citing_title":"Jailbreak Attacks and Defenses Against Large Language Models: A Survey","ref_index":82,"is_internal_anchor":false},{"citing_arxiv_id":"2604.28157","citing_title":"FlashRT: Towards Computationally and Memory Efficient Red-Teaming for Prompt Injection and Knowledge Corruption","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08876","citing_title":"OTora: A Unified Red Teaming Framework for Reasoning-Level Denial-of-Service in LLM Agents","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05058","citing_title":"SoK: Robustness in Large Language Models against Jailbreak Attacks","ref_index":73,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ","json":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ.json","graph_json":"https://pith.science/api/pith-number/YNFBNSISMWIYBMDYY54LDWF7QZ/graph.json","events_json":"https://pith.science/api/pith-number/YNFBNSISMWIYBMDYY54LDWF7QZ/events.json","paper":"https://pith.science/paper/YNFBNSIS"},"agent_actions":{"view_html":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ","download_json":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ.json","view_paper":"https://pith.science/paper/YNFBNSIS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.09674&json=true","fetch_graph":"https://pith.science/api/pith-number/YNFBNSISMWIYBMDYY54LDWF7QZ/graph.json","fetch_events":"https://pith.science/api/pith-number/YNFBNSISMWIYBMDYY54LDWF7QZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ/action/storage_attestation","attest_author":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ/action/author_attestation","sign_citation":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ/action/citation_signature","submit_replication":"https://pith.science/pith/YNFBNSISMWIYBMDYY54LDWF7QZ/action/replication_record"}},"created_at":"2026-07-05T07:45:34.738114+00:00","updated_at":"2026-07-05T07:45:34.738114+00:00"}