{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BQSJY5I4RBKKFK7AMVC62MIHOM","short_pith_number":"pith:BQSJY5I4","schema_version":"1.0","canonical_sha256":"0c249c751c8854a2abe06545ed31077326f0a7677351f944408f89f4ce04f371","source":{"kind":"arxiv","id":"2409.11276","version":1},"attestation_state":"computed","paper":{"title":"Hackphyr: A Local Fine-Tuned LLM Agent for Network Security Environments","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Carlos Catania, Maria Rigaki, Sebastian Garcia","submitted_at":"2024-09-17T15:28:25Z","abstract_excerpt":"Large Language Models (LLMs) have shown remarkable potential across various domains, including cybersecurity. Using commercial cloud-based LLMs may be undesirable due to privacy concerns, costs, and network connectivity constraints. In this paper, we present Hackphyr, a locally fine-tuned LLM to be used as a red-team agent within network security environments. Our fine-tuned 7 billion parameter model can run on a single GPU card and achieves performance comparable with much larger and more powerful commercial models such as GPT-4. Hackphyr clearly outperforms other models, including GPT-3.5-tu"},"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":"2409.11276","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-09-17T15:28:25Z","cross_cats_sorted":[],"title_canon_sha256":"9c1a3cc31ad81d1a1662a3efb67a70d75b5a04012b3153bbe40e597f9b6ddb27","abstract_canon_sha256":"1ba81f24636f5bddabdb0b6be8586c9be3268ab251eb2749ce2d7d501036e9c0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:08:15.905647Z","signature_b64":"YrkquU7UVLHEJsBoDc8CsQ4Bbnpk91CKsGYpd7XIozaOlBE3HDMgKlAuKhWzYcXgOfbxiY5sgHR7C9TULbOnAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0c249c751c8854a2abe06545ed31077326f0a7677351f944408f89f4ce04f371","last_reissued_at":"2026-07-05T09:08:15.905199Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:08:15.905199Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hackphyr: A Local Fine-Tuned LLM Agent for Network Security Environments","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Carlos Catania, Maria Rigaki, Sebastian Garcia","submitted_at":"2024-09-17T15:28:25Z","abstract_excerpt":"Large Language Models (LLMs) have shown remarkable potential across various domains, including cybersecurity. Using commercial cloud-based LLMs may be undesirable due to privacy concerns, costs, and network connectivity constraints. In this paper, we present Hackphyr, a locally fine-tuned LLM to be used as a red-team agent within network security environments. Our fine-tuned 7 billion parameter model can run on a single GPU card and achieves performance comparable with much larger and more powerful commercial models such as GPT-4. Hackphyr clearly outperforms other models, including GPT-3.5-tu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.11276","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/2409.11276/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":"2409.11276","created_at":"2026-07-05T09:08:15.905263+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.11276v1","created_at":"2026-07-05T09:08:15.905263+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.11276","created_at":"2026-07-05T09:08:15.905263+00:00"},{"alias_kind":"pith_short_12","alias_value":"BQSJY5I4RBKK","created_at":"2026-07-05T09:08:15.905263+00:00"},{"alias_kind":"pith_short_16","alias_value":"BQSJY5I4RBKKFK7A","created_at":"2026-07-05T09:08:15.905263+00:00"},{"alias_kind":"pith_short_8","alias_value":"BQSJY5I4","created_at":"2026-07-05T09:08:15.905263+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28479","citing_title":"Decomposing Memorization Reduction in Privacy-Preserving Fine-Tuning of SLMs for CSIRTs","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28146","citing_title":"Cybersecurity AI (CAI) Dataset","ref_index":47,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM","json":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM.json","graph_json":"https://pith.science/api/pith-number/BQSJY5I4RBKKFK7AMVC62MIHOM/graph.json","events_json":"https://pith.science/api/pith-number/BQSJY5I4RBKKFK7AMVC62MIHOM/events.json","paper":"https://pith.science/paper/BQSJY5I4"},"agent_actions":{"view_html":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM","download_json":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM.json","view_paper":"https://pith.science/paper/BQSJY5I4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.11276&json=true","fetch_graph":"https://pith.science/api/pith-number/BQSJY5I4RBKKFK7AMVC62MIHOM/graph.json","fetch_events":"https://pith.science/api/pith-number/BQSJY5I4RBKKFK7AMVC62MIHOM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM/action/storage_attestation","attest_author":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM/action/author_attestation","sign_citation":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM/action/citation_signature","submit_replication":"https://pith.science/pith/BQSJY5I4RBKKFK7AMVC62MIHOM/action/replication_record"}},"created_at":"2026-07-05T09:08:15.905263+00:00","updated_at":"2026-07-05T09:08:15.905263+00:00"}