{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QCU3G3IB45MD64HTSTJ5XZ33IA","short_pith_number":"pith:QCU3G3IB","schema_version":"1.0","canonical_sha256":"80a9b36d01e7583f70f394d3dbe77b403e33878afc4a3c4ba1c3cce9f3b6bbc3","source":{"kind":"arxiv","id":"2607.20581","version":1},"attestation_state":"computed","paper":{"title":"Geometric Configurations of Perturbed Jailbreak Prompts","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Andres Algaba, Lynn Delcon, Vincent Ginis","submitted_at":"2026-07-22T11:52:43Z","abstract_excerpt":"Perturbation techniques that turn unsuccessful jailbreak prompts into successful ones are continuously evolving, constituting a major security threat to LLM safety. In this paper, we investigate the internal representations of such string-level perturbed jailbreak inputs in the small weight models of the Qwen-2.5-1.5B/-3B/-7B-Instruct and Llama-3.2-1B/-3B/-3.1-8B-Instruct families. We select two representation spaces: the last-layer-last-token embedding space and the top-50 next-token probability space. The former space separates prompts based on their spelling and format, while the latter spa"},"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":"2607.20581","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2026-07-22T11:52:43Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"2a1ecf4acb5ac7c748528d9d854b27cbb19ecf93f0dc32d16667e21fedecb0eb","abstract_canon_sha256":"c6093263fd591c085a46cc6d96739062360ebc7b0f89ad41ee9f66d6e5683f65"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-24T00:23:23.704863Z","signature_b64":"Ori69sa9m/4TcHX3vF6d+JTtq2VAfdqWjmDugug5N4a8QcSFuzP5TsFQ/YAUbpAMIIWFG4b3V+py0hLVzkjRAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"80a9b36d01e7583f70f394d3dbe77b403e33878afc4a3c4ba1c3cce9f3b6bbc3","last_reissued_at":"2026-07-24T00:23:23.703933Z","signature_status":"signed_v1","first_computed_at":"2026-07-24T00:23:23.703933Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Geometric Configurations of Perturbed Jailbreak Prompts","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Andres Algaba, Lynn Delcon, Vincent Ginis","submitted_at":"2026-07-22T11:52:43Z","abstract_excerpt":"Perturbation techniques that turn unsuccessful jailbreak prompts into successful ones are continuously evolving, constituting a major security threat to LLM safety. In this paper, we investigate the internal representations of such string-level perturbed jailbreak inputs in the small weight models of the Qwen-2.5-1.5B/-3B/-7B-Instruct and Llama-3.2-1B/-3B/-3.1-8B-Instruct families. We select two representation spaces: the last-layer-last-token embedding space and the top-50 next-token probability space. The former space separates prompts based on their spelling and format, while the latter spa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.20581","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/2607.20581/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":"2607.20581","created_at":"2026-07-24T00:23:23.704378+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.20581v1","created_at":"2026-07-24T00:23:23.704378+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.20581","created_at":"2026-07-24T00:23:23.704378+00:00"},{"alias_kind":"pith_short_12","alias_value":"QCU3G3IB45MD","created_at":"2026-07-24T00:23:23.704378+00:00"},{"alias_kind":"pith_short_16","alias_value":"QCU3G3IB45MD64HT","created_at":"2026-07-24T00:23:23.704378+00:00"},{"alias_kind":"pith_short_8","alias_value":"QCU3G3IB","created_at":"2026-07-24T00:23:23.704378+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA","json":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA.json","graph_json":"https://pith.science/api/pith-number/QCU3G3IB45MD64HTSTJ5XZ33IA/graph.json","events_json":"https://pith.science/api/pith-number/QCU3G3IB45MD64HTSTJ5XZ33IA/events.json","paper":"https://pith.science/paper/QCU3G3IB"},"agent_actions":{"view_html":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA","download_json":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA.json","view_paper":"https://pith.science/paper/QCU3G3IB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.20581&json=true","fetch_graph":"https://pith.science/api/pith-number/QCU3G3IB45MD64HTSTJ5XZ33IA/graph.json","fetch_events":"https://pith.science/api/pith-number/QCU3G3IB45MD64HTSTJ5XZ33IA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA/action/storage_attestation","attest_author":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA/action/author_attestation","sign_citation":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA/action/citation_signature","submit_replication":"https://pith.science/pith/QCU3G3IB45MD64HTSTJ5XZ33IA/action/replication_record"}},"created_at":"2026-07-24T00:23:23.704378+00:00","updated_at":"2026-07-24T00:23:23.704378+00:00"}