{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:T4XA4GCBW4UW6NGD3AUZW2FIHK","short_pith_number":"pith:T4XA4GCB","canonical_record":{"source":{"id":"2502.13603","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-19T10:33:18Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"d0ac2496a0675454b7fa54eed96877bb43c601edf0adce71777f76dde535a0db","abstract_canon_sha256":"04657c946fd6ae1cfe24dcefa05472ff04614159e5950e8f26e05d5d73328750"},"schema_version":"1.0"},"canonical_sha256":"9f2e0e1841b7296f34c3d8299b68a83a8ba362fdead45008be8a457f16ba143e","source":{"kind":"arxiv","id":"2502.13603","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.13603","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"arxiv_version","alias_value":"2502.13603v2","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13603","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_12","alias_value":"T4XA4GCBW4UW","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_16","alias_value":"T4XA4GCBW4UW6NGD","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_8","alias_value":"T4XA4GCB","created_at":"2026-07-05T10:19:39Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:T4XA4GCBW4UW6NGD3AUZW2FIHK","target":"record","payload":{"canonical_record":{"source":{"id":"2502.13603","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-19T10:33:18Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"d0ac2496a0675454b7fa54eed96877bb43c601edf0adce71777f76dde535a0db","abstract_canon_sha256":"04657c946fd6ae1cfe24dcefa05472ff04614159e5950e8f26e05d5d73328750"},"schema_version":"1.0"},"canonical_sha256":"9f2e0e1841b7296f34c3d8299b68a83a8ba362fdead45008be8a457f16ba143e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:19:39.573071Z","signature_b64":"KSrere6gx1rd9ZdYsMKf5MHuNeiBLcn8CU0SW7laT/yqlltc4SyXtUsmRlpHa7LEytaeOo/p+kMehlGWmYFzDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9f2e0e1841b7296f34c3d8299b68a83a8ba362fdead45008be8a457f16ba143e","last_reissued_at":"2026-07-05T10:19:39.572575Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:19:39.572575Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.13603","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:19:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"A7pPoDgQk5a2vjgRpXRpWOw9A1gdXv/K9I0P1SjswLb0q9cTdVtFQkGYttKn9ytRU1BneLAiUII12NTJc6aYBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T04:05:57.907882Z"},"content_sha256":"062d20f27f1650fec094584325e0827b17c2273592bc50d798f7f6817c9caf0d","schema_version":"1.0","event_id":"sha256:062d20f27f1650fec094584325e0827b17c2273592bc50d798f7f6817c9caf0d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:T4XA4GCBW4UW6NGD3AUZW2FIHK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Efficient Safety Retrofitting Against Jailbreaking for LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Adrian Tormos, Anna Arias-Duart, Ashwin Kumar Gururajan, Daniel Hinjos, Dario Garcia-Gasulla, Maria Eugenia Cardello, Oscar Molina-Sedano","submitted_at":"2025-02-19T10:33:18Z","abstract_excerpt":"Direct Preference Optimization (DPO) is an efficient alignment technique that steers LLMs towards preferable outputs by training on preference data, bypassing the need for explicit reward models. Its simplicity enables easy adaptation to various domains and safety requirements. This paper examines DPO's effectiveness in model safety against jailbreaking attacks while minimizing data requirements and training costs. We introduce Egida, a dataset expanded from multiple sources, which includes 27 different safety topics and 18 different attack styles, complemented with synthetic and human labels."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13603","kind":"arxiv","version":2},"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/2502.13603/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":"2026-07-05T10:19:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0xJn4DlfBhW3k6Ofais69+hcGAsEkT82LhARL0NcbB6n0XabtgshqCaOWhuAPT17bEi6wnULszXrMxTSs+/CBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T04:05:57.908882Z"},"content_sha256":"66837396dd50b8e50b3ee4fc20c8653011b44d990171f947465f7bb510831c0f","schema_version":"1.0","event_id":"sha256:66837396dd50b8e50b3ee4fc20c8653011b44d990171f947465f7bb510831c0f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/T4XA4GCBW4UW6NGD3AUZW2FIHK/bundle.json","state_url":"https://pith.science/pith/T4XA4GCBW4UW6NGD3AUZW2FIHK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/T4XA4GCBW4UW6NGD3AUZW2FIHK/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":"2026-08-21T04:05:57Z","links":{"resolver":"https://pith.science/pith/T4XA4GCBW4UW6NGD3AUZW2FIHK","bundle":"https://pith.science/pith/T4XA4GCBW4UW6NGD3AUZW2FIHK/bundle.json","state":"https://pith.science/pith/T4XA4GCBW4UW6NGD3AUZW2FIHK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/T4XA4GCBW4UW6NGD3AUZW2FIHK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:T4XA4GCBW4UW6NGD3AUZW2FIHK","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":"04657c946fd6ae1cfe24dcefa05472ff04614159e5950e8f26e05d5d73328750","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-19T10:33:18Z","title_canon_sha256":"d0ac2496a0675454b7fa54eed96877bb43c601edf0adce71777f76dde535a0db"},"schema_version":"1.0","source":{"id":"2502.13603","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.13603","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"arxiv_version","alias_value":"2502.13603v2","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13603","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_12","alias_value":"T4XA4GCBW4UW","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_16","alias_value":"T4XA4GCBW4UW6NGD","created_at":"2026-07-05T10:19:39Z"},{"alias_kind":"pith_short_8","alias_value":"T4XA4GCB","created_at":"2026-07-05T10:19:39Z"}],"graph_snapshots":[{"event_id":"sha256:66837396dd50b8e50b3ee4fc20c8653011b44d990171f947465f7bb510831c0f","target":"graph","created_at":"2026-07-05T10:19:39Z","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/2502.13603/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Direct Preference Optimization (DPO) is an efficient alignment technique that steers LLMs towards preferable outputs by training on preference data, bypassing the need for explicit reward models. Its simplicity enables easy adaptation to various domains and safety requirements. This paper examines DPO's effectiveness in model safety against jailbreaking attacks while minimizing data requirements and training costs. We introduce Egida, a dataset expanded from multiple sources, which includes 27 different safety topics and 18 different attack styles, complemented with synthetic and human labels.","authors_text":"Adrian Tormos, Anna Arias-Duart, Ashwin Kumar Gururajan, Daniel Hinjos, Dario Garcia-Gasulla, Maria Eugenia Cardello, Oscar Molina-Sedano","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-19T10:33:18Z","title":"Efficient Safety Retrofitting Against Jailbreaking for LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13603","kind":"arxiv","version":2},"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":"sha256:062d20f27f1650fec094584325e0827b17c2273592bc50d798f7f6817c9caf0d","target":"record","created_at":"2026-07-05T10:19:39Z","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":"04657c946fd6ae1cfe24dcefa05472ff04614159e5950e8f26e05d5d73328750","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-19T10:33:18Z","title_canon_sha256":"d0ac2496a0675454b7fa54eed96877bb43c601edf0adce71777f76dde535a0db"},"schema_version":"1.0","source":{"id":"2502.13603","kind":"arxiv","version":2}},"canonical_sha256":"9f2e0e1841b7296f34c3d8299b68a83a8ba362fdead45008be8a457f16ba143e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9f2e0e1841b7296f34c3d8299b68a83a8ba362fdead45008be8a457f16ba143e","first_computed_at":"2026-07-05T10:19:39.572575Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:19:39.572575Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KSrere6gx1rd9ZdYsMKf5MHuNeiBLcn8CU0SW7laT/yqlltc4SyXtUsmRlpHa7LEytaeOo/p+kMehlGWmYFzDA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:19:39.573071Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.13603","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:062d20f27f1650fec094584325e0827b17c2273592bc50d798f7f6817c9caf0d","sha256:66837396dd50b8e50b3ee4fc20c8653011b44d990171f947465f7bb510831c0f"],"state_sha256":"bd4e728de54a052ccf00008b52914dd013f4b3631205bd349ce8c2a34ca5b85b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"G/RveaQztDwjd1sqAg6c6cDv1zPnSt4HZpEgylEJwwKPIbKd4fuwamyL930BWv7CWfAFulcw0ebf8qRSHGIUAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-21T04:05:57.914868Z","bundle_sha256":"a9d5a5d5dfbd210d4b2e80df39a7fbe3de144fe49e9025fc0e0a661f32552ab9"}}