{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:A7424GN73OGWJ4SDYJ7WSVQAHG","short_pith_number":"pith:A7424GN7","canonical_record":{"source":{"id":"2401.10862","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-19T18:05:34Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"title_canon_sha256":"a46d96e8c78c8601a0336c54b6b9387edd1c4d001ebd597f376ea294e3c7c3b2","abstract_canon_sha256":"8a612e745d48013c61617b7d5854160940f4ddf6e71887f2c14cf6c98098769a"},"schema_version":"1.0"},"canonical_sha256":"07f9ae19bfdb8d64f243c27f6956003997d945589156acade690c75e0513c434","source":{"kind":"arxiv","id":"2401.10862","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.10862","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"arxiv_version","alias_value":"2401.10862v3","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.10862","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"pith_short_12","alias_value":"A7424GN73OGW","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"pith_short_16","alias_value":"A7424GN73OGWJ4SD","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"pith_short_8","alias_value":"A7424GN7","created_at":"2026-07-05T09:28:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:A7424GN73OGWJ4SDYJ7WSVQAHG","target":"record","payload":{"canonical_record":{"source":{"id":"2401.10862","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-19T18:05:34Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"title_canon_sha256":"a46d96e8c78c8601a0336c54b6b9387edd1c4d001ebd597f376ea294e3c7c3b2","abstract_canon_sha256":"8a612e745d48013c61617b7d5854160940f4ddf6e71887f2c14cf6c98098769a"},"schema_version":"1.0"},"canonical_sha256":"07f9ae19bfdb8d64f243c27f6956003997d945589156acade690c75e0513c434","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:28:47.210406Z","signature_b64":"D37EG0aZ8pivuVCNrsr8EhxX+jC7L8Fy0YMfAVws/ZVwp4ZHTrv065ZbybKQdJmavePRzKf2YVERz9sDX1KWDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"07f9ae19bfdb8d64f243c27f6956003997d945589156acade690c75e0513c434","last_reissued_at":"2026-07-05T09:28:47.209912Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:28:47.209912Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2401.10862","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:28:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kcPDv3Ec3xtjd0vmBUX9bkRvhTnKKe3k9MZ2QMxsvtejL5tDSYNPCX8eqspCsQ7vqhIO4Tb1EddcqaMBdNRkCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T16:15:26.809252Z"},"content_sha256":"6a996e4eab20ab581d8cf22b2e08e02db0a66da889f7e06c072fc2ca3c37f2a0","schema_version":"1.0","event_id":"sha256:6a996e4eab20ab581d8cf22b2e08e02db0a66da889f7e06c072fc2ca3c37f2a0"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:A7424GN73OGWJ4SDYJ7WSVQAHG","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Pruning for Protection: Increasing Jailbreak Resistance in Aligned LLMs Without Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CR"],"primary_cat":"cs.LG","authors_text":"Adib Hasan, Alex Wang, Ileana Rugina","submitted_at":"2024-01-19T18:05:34Z","abstract_excerpt":"This paper investigates the impact of model compression on the way Large Language Models (LLMs) process prompts, particularly concerning jailbreak resistance. We show that moderate WANDA pruning can enhance resistance to jailbreaking attacks without fine-tuning, while maintaining performance on standard benchmarks. To systematically evaluate this safety enhancement, we introduce a dataset of 225 harmful tasks across five categories. Our analysis of LLaMA-2 Chat, Vicuna 1.3, and Mistral Instruct v0.2 reveals that pruning benefits correlate with initial model safety levels. We interpret these re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.10862","kind":"arxiv","version":3},"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/2401.10862/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":"2026-07-05T09:28:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"g1EuLMKcUeR9GzZa7L0Exxr2MiDLL9e0RbtivzxRH0pnr/5i1od+m7P5GqCcd8k16/4fmPW6g17f/VSQVHlGDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T16:15:26.810419Z"},"content_sha256":"c5e1dd90f9497a1855968fcc620f64204815eb0dcf755dc8e69b5ff60584e712","schema_version":"1.0","event_id":"sha256:c5e1dd90f9497a1855968fcc620f64204815eb0dcf755dc8e69b5ff60584e712"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/A7424GN73OGWJ4SDYJ7WSVQAHG/bundle.json","state_url":"https://pith.science/pith/A7424GN73OGWJ4SDYJ7WSVQAHG/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/A7424GN73OGWJ4SDYJ7WSVQAHG/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":"2026-08-04T16:15:26Z","links":{"resolver":"https://pith.science/pith/A7424GN73OGWJ4SDYJ7WSVQAHG","bundle":"https://pith.science/pith/A7424GN73OGWJ4SDYJ7WSVQAHG/bundle.json","state":"https://pith.science/pith/A7424GN73OGWJ4SDYJ7WSVQAHG/state.json","well_known_bundle":"https://pith.science/.well-known/pith/A7424GN73OGWJ4SDYJ7WSVQAHG/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:A7424GN73OGWJ4SDYJ7WSVQAHG","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":"8a612e745d48013c61617b7d5854160940f4ddf6e71887f2c14cf6c98098769a","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-19T18:05:34Z","title_canon_sha256":"a46d96e8c78c8601a0336c54b6b9387edd1c4d001ebd597f376ea294e3c7c3b2"},"schema_version":"1.0","source":{"id":"2401.10862","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.10862","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"arxiv_version","alias_value":"2401.10862v3","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.10862","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"pith_short_12","alias_value":"A7424GN73OGW","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"pith_short_16","alias_value":"A7424GN73OGWJ4SD","created_at":"2026-07-05T09:28:47Z"},{"alias_kind":"pith_short_8","alias_value":"A7424GN7","created_at":"2026-07-05T09:28:47Z"}],"graph_snapshots":[{"event_id":"sha256:c5e1dd90f9497a1855968fcc620f64204815eb0dcf755dc8e69b5ff60584e712","target":"graph","created_at":"2026-07-05T09:28:47Z","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/2401.10862/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"This paper investigates the impact of model compression on the way Large Language Models (LLMs) process prompts, particularly concerning jailbreak resistance. We show that moderate WANDA pruning can enhance resistance to jailbreaking attacks without fine-tuning, while maintaining performance on standard benchmarks. To systematically evaluate this safety enhancement, we introduce a dataset of 225 harmful tasks across five categories. Our analysis of LLaMA-2 Chat, Vicuna 1.3, and Mistral Instruct v0.2 reveals that pruning benefits correlate with initial model safety levels. We interpret these re","authors_text":"Adib Hasan, Alex Wang, Ileana Rugina","cross_cats":["cs.AI","cs.CL","cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-19T18:05:34Z","title":"Pruning for Protection: Increasing Jailbreak Resistance in Aligned LLMs Without Fine-Tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.10862","kind":"arxiv","version":3},"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":"sha256:6a996e4eab20ab581d8cf22b2e08e02db0a66da889f7e06c072fc2ca3c37f2a0","target":"record","created_at":"2026-07-05T09:28:47Z","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":"8a612e745d48013c61617b7d5854160940f4ddf6e71887f2c14cf6c98098769a","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-19T18:05:34Z","title_canon_sha256":"a46d96e8c78c8601a0336c54b6b9387edd1c4d001ebd597f376ea294e3c7c3b2"},"schema_version":"1.0","source":{"id":"2401.10862","kind":"arxiv","version":3}},"canonical_sha256":"07f9ae19bfdb8d64f243c27f6956003997d945589156acade690c75e0513c434","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"07f9ae19bfdb8d64f243c27f6956003997d945589156acade690c75e0513c434","first_computed_at":"2026-07-05T09:28:47.209912Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:28:47.209912Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"D37EG0aZ8pivuVCNrsr8EhxX+jC7L8Fy0YMfAVws/ZVwp4ZHTrv065ZbybKQdJmavePRzKf2YVERz9sDX1KWDg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:28:47.210406Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.10862","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6a996e4eab20ab581d8cf22b2e08e02db0a66da889f7e06c072fc2ca3c37f2a0","sha256:c5e1dd90f9497a1855968fcc620f64204815eb0dcf755dc8e69b5ff60584e712"],"state_sha256":"981ab3e5e1da82c47ecc8c448637d4e4ae6030dc351cae584f8e4cdb74f660ed"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ze2q2bD2NG5E6cxNKRWJ6YimkjCZuU0dOZ50SHDM7Wkla0DiFYzgUrdL/EjAlM0iuV4oSRm3VoLWvAchh72GBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T16:15:26.815640Z","bundle_sha256":"34abc75ecee2b523d6c13fae877d8d3c595c91f851090ef352b7866c720c1356"}}