{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5IMD5MXTWQINNPCOWM5SENUCC5","short_pith_number":"pith:5IMD5MXT","schema_version":"1.0","canonical_sha256":"ea183eb2f3b410d6bc4eb33b22368217610d4c793b25f647512dcff072147ae7","source":{"kind":"arxiv","id":"2506.12274","version":1},"attestation_state":"computed","paper":{"title":"InfoFlood: Jailbreaking Large Language Models with Information Overload","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CR","authors_text":"Advait Yadav, Haibo Jin, Haohan Wang, Jun Zhuang, Man Luo","submitted_at":"2025-06-13T23:03:11Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains. However, their potential to generate harmful responses has raised significant societal and regulatory concerns, especially when manipulated by adversarial techniques known as \"jailbreak\" attacks. Existing jailbreak methods typically involve appending carefully crafted prefixes or suffixes to malicious prompts in order to bypass the built-in safety mechanisms of these models.\n  In this work, we identify a new vulnerability in which excessive linguistic complexity can disrupt built-in safety mechanisms"},"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":"2506.12274","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-06-13T23:03:11Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"9f401416e4d5953365fbdc3789dc6ac59e2681f6f0fed179f90cdbb2c6b24f05","abstract_canon_sha256":"7f6f875e452c5039ede15ab3b94e71fb488488c76c94ced0881207cca7df2166"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:21:50.633173Z","signature_b64":"qypTkoPJIuyOK5uEtP00voymZdqv6z7FuwER8lEG9wVvd0Jm25pla/Gipc98NAZezf5ksSnDMMrxaUeCJgdCCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ea183eb2f3b410d6bc4eb33b22368217610d4c793b25f647512dcff072147ae7","last_reissued_at":"2026-07-05T11:21:50.632737Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:21:50.632737Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"InfoFlood: Jailbreaking Large Language Models with Information Overload","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CR","authors_text":"Advait Yadav, Haibo Jin, Haohan Wang, Jun Zhuang, Man Luo","submitted_at":"2025-06-13T23:03:11Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains. However, their potential to generate harmful responses has raised significant societal and regulatory concerns, especially when manipulated by adversarial techniques known as \"jailbreak\" attacks. Existing jailbreak methods typically involve appending carefully crafted prefixes or suffixes to malicious prompts in order to bypass the built-in safety mechanisms of these models.\n  In this work, we identify a new vulnerability in which excessive linguistic complexity can disrupt built-in safety mechanisms"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.12274","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/2506.12274/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":"2506.12274","created_at":"2026-07-05T11:21:50.632793+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.12274v1","created_at":"2026-07-05T11:21:50.632793+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.12274","created_at":"2026-07-05T11:21:50.632793+00:00"},{"alias_kind":"pith_short_12","alias_value":"5IMD5MXTWQIN","created_at":"2026-07-05T11:21:50.632793+00:00"},{"alias_kind":"pith_short_16","alias_value":"5IMD5MXTWQINNPCO","created_at":"2026-07-05T11:21:50.632793+00:00"},{"alias_kind":"pith_short_8","alias_value":"5IMD5MXT","created_at":"2026-07-05T11:21:50.632793+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.10546","citing_title":"Learning to Conceal Risk: Controllable Multi-turn Red Teaming for LLMs in the Financial Domain","ref_index":40,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5","json":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5.json","graph_json":"https://pith.science/api/pith-number/5IMD5MXTWQINNPCOWM5SENUCC5/graph.json","events_json":"https://pith.science/api/pith-number/5IMD5MXTWQINNPCOWM5SENUCC5/events.json","paper":"https://pith.science/paper/5IMD5MXT"},"agent_actions":{"view_html":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5","download_json":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5.json","view_paper":"https://pith.science/paper/5IMD5MXT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.12274&json=true","fetch_graph":"https://pith.science/api/pith-number/5IMD5MXTWQINNPCOWM5SENUCC5/graph.json","fetch_events":"https://pith.science/api/pith-number/5IMD5MXTWQINNPCOWM5SENUCC5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5/action/storage_attestation","attest_author":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5/action/author_attestation","sign_citation":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5/action/citation_signature","submit_replication":"https://pith.science/pith/5IMD5MXTWQINNPCOWM5SENUCC5/action/replication_record"}},"created_at":"2026-07-05T11:21:50.632793+00:00","updated_at":"2026-07-05T11:21:50.632793+00:00"}