{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:PPM52GXQQ6YE44TX74W3C2OJTS","short_pith_number":"pith:PPM52GXQ","canonical_record":{"source":{"id":"2507.14202","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-14T17:41:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ca7b0f7bcb45ba55c1c6a65132f51c0a1b2634a372e8118e6fb6f7945cef684a","abstract_canon_sha256":"22fa65229e8ee36f02bae7802207d75cf40f514d60c6e5a69ee636fc23ebeb8c"},"schema_version":"1.0"},"canonical_sha256":"7bd9dd1af087b04e7277ff2db169c99c931ebe0dd430b1c09c1d6b3afcad364b","source":{"kind":"arxiv","id":"2507.14202","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.14202","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"arxiv_version","alias_value":"2507.14202v1","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.14202","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"pith_short_12","alias_value":"PPM52GXQQ6YE","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"pith_short_16","alias_value":"PPM52GXQQ6YE44TX","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"pith_short_8","alias_value":"PPM52GXQ","created_at":"2026-07-05T11:39:30Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:PPM52GXQQ6YE44TX74W3C2OJTS","target":"record","payload":{"canonical_record":{"source":{"id":"2507.14202","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-14T17:41:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ca7b0f7bcb45ba55c1c6a65132f51c0a1b2634a372e8118e6fb6f7945cef684a","abstract_canon_sha256":"22fa65229e8ee36f02bae7802207d75cf40f514d60c6e5a69ee636fc23ebeb8c"},"schema_version":"1.0"},"canonical_sha256":"7bd9dd1af087b04e7277ff2db169c99c931ebe0dd430b1c09c1d6b3afcad364b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:39:30.674925Z","signature_b64":"XXr7ptL2oAUnF2sjqn6PL4wjGco0Dx8bzsE619DmJ4FYeIZoYiawQnoWJMUB7LPId1t4c2IbFgIaTk3FpXzJDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7bd9dd1af087b04e7277ff2db169c99c931ebe0dd430b1c09c1d6b3afcad364b","last_reissued_at":"2026-07-05T11:39:30.674517Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:39:30.674517Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2507.14202","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:39:30Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cdd77Vr8KQuz4RwC+0iQmYSCbqwu4E9F8Ak6XF8hAgkCbh1mwmImMl8cInQuf1QH+KRJybn/KHkJ0qgxOS5uBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T09:00:10.389507Z"},"content_sha256":"5bfcaa3b8dcaf3acea60ba12f65f882f1da3be15621297620e53078e90e0f8f2","schema_version":"1.0","event_id":"sha256:5bfcaa3b8dcaf3acea60ba12f65f882f1da3be15621297620e53078e90e0f8f2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:PPM52GXQQ6YE44TX74W3C2OJTS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"PRM-Free Security Alignment of Large Models via Red Teaming and Adversarial Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Pengfei Du","submitted_at":"2025-07-14T17:41:12Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse applications, yet they pose significant security risks that threaten their safe deployment in critical domains. Current security alignment methodologies predominantly rely on Process Reward Models (PRMs) to evaluate intermediate reasoning steps, introducing substantial computational overhead and scalability constraints. This paper presents a novel PRM-free security alignment framework that leverages automated red teaming and adversarial training to achieve robust security guarantees while maintaining computat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.14202","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/2507.14202/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:39:30Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Jv98rRRWgyV1D/8DzjwSBitVn0Az8me/52XTDyIn1BEi+rXYT2LvhyAvsJuE3v83umekzleghFVZ9c0mqTlpDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T09:00:10.390605Z"},"content_sha256":"608b5483060d2cdf54863269bf08ac15dd374badafbb8b139bd8f2d34c1b3a00","schema_version":"1.0","event_id":"sha256:608b5483060d2cdf54863269bf08ac15dd374badafbb8b139bd8f2d34c1b3a00"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/PPM52GXQQ6YE44TX74W3C2OJTS/bundle.json","state_url":"https://pith.science/pith/PPM52GXQQ6YE44TX74W3C2OJTS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/PPM52GXQQ6YE44TX74W3C2OJTS/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-11T09:00:10Z","links":{"resolver":"https://pith.science/pith/PPM52GXQQ6YE44TX74W3C2OJTS","bundle":"https://pith.science/pith/PPM52GXQQ6YE44TX74W3C2OJTS/bundle.json","state":"https://pith.science/pith/PPM52GXQQ6YE44TX74W3C2OJTS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/PPM52GXQQ6YE44TX74W3C2OJTS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:PPM52GXQQ6YE44TX74W3C2OJTS","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"22fa65229e8ee36f02bae7802207d75cf40f514d60c6e5a69ee636fc23ebeb8c","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-14T17:41:12Z","title_canon_sha256":"ca7b0f7bcb45ba55c1c6a65132f51c0a1b2634a372e8118e6fb6f7945cef684a"},"schema_version":"1.0","source":{"id":"2507.14202","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.14202","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"arxiv_version","alias_value":"2507.14202v1","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.14202","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"pith_short_12","alias_value":"PPM52GXQQ6YE","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"pith_short_16","alias_value":"PPM52GXQQ6YE44TX","created_at":"2026-07-05T11:39:30Z"},{"alias_kind":"pith_short_8","alias_value":"PPM52GXQ","created_at":"2026-07-05T11:39:30Z"}],"graph_snapshots":[{"event_id":"sha256:608b5483060d2cdf54863269bf08ac15dd374badafbb8b139bd8f2d34c1b3a00","target":"graph","created_at":"2026-07-05T11:39:30Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2507.14202/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse applications, yet they pose significant security risks that threaten their safe deployment in critical domains. Current security alignment methodologies predominantly rely on Process Reward Models (PRMs) to evaluate intermediate reasoning steps, introducing substantial computational overhead and scalability constraints. This paper presents a novel PRM-free security alignment framework that leverages automated red teaming and adversarial training to achieve robust security guarantees while maintaining computat","authors_text":"Pengfei Du","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-14T17:41:12Z","title":"PRM-Free Security Alignment of Large Models via Red Teaming and Adversarial Training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.14202","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:5bfcaa3b8dcaf3acea60ba12f65f882f1da3be15621297620e53078e90e0f8f2","target":"record","created_at":"2026-07-05T11:39:30Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"22fa65229e8ee36f02bae7802207d75cf40f514d60c6e5a69ee636fc23ebeb8c","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-14T17:41:12Z","title_canon_sha256":"ca7b0f7bcb45ba55c1c6a65132f51c0a1b2634a372e8118e6fb6f7945cef684a"},"schema_version":"1.0","source":{"id":"2507.14202","kind":"arxiv","version":1}},"canonical_sha256":"7bd9dd1af087b04e7277ff2db169c99c931ebe0dd430b1c09c1d6b3afcad364b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7bd9dd1af087b04e7277ff2db169c99c931ebe0dd430b1c09c1d6b3afcad364b","first_computed_at":"2026-07-05T11:39:30.674517Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:39:30.674517Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"XXr7ptL2oAUnF2sjqn6PL4wjGco0Dx8bzsE619DmJ4FYeIZoYiawQnoWJMUB7LPId1t4c2IbFgIaTk3FpXzJDw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:39:30.674925Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.14202","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5bfcaa3b8dcaf3acea60ba12f65f882f1da3be15621297620e53078e90e0f8f2","sha256:608b5483060d2cdf54863269bf08ac15dd374badafbb8b139bd8f2d34c1b3a00"],"state_sha256":"3b38f78adcb3ad7a8dae7929328968328978d3c9dbb640d51ca2bfe37f42f212"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"j9/fipkTlqaY+Gy5Bjv982181f77wdkdWmyaKzN+SQG2WJQ04m8v73bzPzvEL49A7wWiRPP/92ed/40lasOGBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-11T09:00:10.399756Z","bundle_sha256":"4db959435233070dabb9a3eab51d7956e2dc85d7b8b5783957553323412dcc65"}}