{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:QJJTFYK22COC6AASI35HUD3DQS","short_pith_number":"pith:QJJTFYK2","canonical_record":{"source":{"id":"2411.02785","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-05T03:51:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9f46a5dbbf4e2638b7de4c39049a77c2ad0f4d6006de1ca80f8fa406c84875cc","abstract_canon_sha256":"50a7d6c9d4359568beca4d311e7bf34369ab32af446a7cf33de931fb2580571f"},"schema_version":"1.0"},"canonical_sha256":"825332e15ad09c2f001246fa7a0f6384a22d11274e48e467046355cc046418a0","source":{"kind":"arxiv","id":"2411.02785","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.02785","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"arxiv_version","alias_value":"2411.02785v2","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.02785","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"pith_short_12","alias_value":"QJJTFYK22COC","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"pith_short_16","alias_value":"QJJTFYK22COC6AAS","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"pith_short_8","alias_value":"QJJTFYK2","created_at":"2026-07-05T09:44:54Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:QJJTFYK22COC6AASI35HUD3DQS","target":"record","payload":{"canonical_record":{"source":{"id":"2411.02785","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-05T03:51:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9f46a5dbbf4e2638b7de4c39049a77c2ad0f4d6006de1ca80f8fa406c84875cc","abstract_canon_sha256":"50a7d6c9d4359568beca4d311e7bf34369ab32af446a7cf33de931fb2580571f"},"schema_version":"1.0"},"canonical_sha256":"825332e15ad09c2f001246fa7a0f6384a22d11274e48e467046355cc046418a0","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:44:54.346043Z","signature_b64":"w5/2SrMz+SsfvwiMhhFZzSyco8IujHHxg4SVq73eyvSFpLRhFAmIef+RGXTsl4Fq/86/0bPqqoWqXyoB03mFAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"825332e15ad09c2f001246fa7a0f6384a22d11274e48e467046355cc046418a0","last_reissued_at":"2026-07-05T09:44:54.345540Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:44:54.345540Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2411.02785","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":"2026-07-05T09:44:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"bHsjC1/9u31Pu6E/O5rZsmursFHJyDGPoaLjFuxqE30EXL2HPin6uSNmFSoqmPiC6UL2bJljaeWOxKhN7x5bCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T13:59:41.504494Z"},"content_sha256":"13fc54d9c83d4b3cbaa9a72498322dc5b25b828ab16d212f2d50deadf7134b1c","schema_version":"1.0","event_id":"sha256:13fc54d9c83d4b3cbaa9a72498322dc5b25b828ab16d212f2d50deadf7134b1c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:QJJTFYK22COC6AASI35HUD3DQS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Stochastic Monkeys at Play: Random Augmentations Cheaply Break LLM Safety Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Gagandeep Singh, Gaokai Zhang, Hangoo Kang, Jason Vega, Junsheng Huang, Minjia Zhang","submitted_at":"2024-11-05T03:51:13Z","abstract_excerpt":"Safety alignment of Large Language Models (LLMs) has recently become a critical objective of model developers. In response, a growing body of work has been investigating how safety alignment can be bypassed through various jailbreaking methods, such as adversarial attacks. However, these jailbreak methods can be rather costly or involve a non-trivial amount of creativity and effort, introducing the assumption that malicious users are high-resource or sophisticated. In this paper, we study how simple random augmentations to the input prompt affect safety alignment effectiveness in state-of-the-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.02785","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/2411.02785/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":"2026-07-05T09:44:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7IBWy4BT2UHoh9kID01RkexDI63nS+YqSAHt5rclDxpWBgO/z42sTo5dKwoni3f+ZXPqKK+vFlyK5XXGImejCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T13:59:41.505506Z"},"content_sha256":"45c0365936fc29464dda897ec8dedb3e60d1a118aad6045de86c30dc18795675","schema_version":"1.0","event_id":"sha256:45c0365936fc29464dda897ec8dedb3e60d1a118aad6045de86c30dc18795675"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/QJJTFYK22COC6AASI35HUD3DQS/bundle.json","state_url":"https://pith.science/pith/QJJTFYK22COC6AASI35HUD3DQS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/QJJTFYK22COC6AASI35HUD3DQS/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":"2026-08-08T13:59:41Z","links":{"resolver":"https://pith.science/pith/QJJTFYK22COC6AASI35HUD3DQS","bundle":"https://pith.science/pith/QJJTFYK22COC6AASI35HUD3DQS/bundle.json","state":"https://pith.science/pith/QJJTFYK22COC6AASI35HUD3DQS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/QJJTFYK22COC6AASI35HUD3DQS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:QJJTFYK22COC6AASI35HUD3DQS","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":"50a7d6c9d4359568beca4d311e7bf34369ab32af446a7cf33de931fb2580571f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-05T03:51:13Z","title_canon_sha256":"9f46a5dbbf4e2638b7de4c39049a77c2ad0f4d6006de1ca80f8fa406c84875cc"},"schema_version":"1.0","source":{"id":"2411.02785","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.02785","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"arxiv_version","alias_value":"2411.02785v2","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.02785","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"pith_short_12","alias_value":"QJJTFYK22COC","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"pith_short_16","alias_value":"QJJTFYK22COC6AAS","created_at":"2026-07-05T09:44:54Z"},{"alias_kind":"pith_short_8","alias_value":"QJJTFYK2","created_at":"2026-07-05T09:44:54Z"}],"graph_snapshots":[{"event_id":"sha256:45c0365936fc29464dda897ec8dedb3e60d1a118aad6045de86c30dc18795675","target":"graph","created_at":"2026-07-05T09:44:54Z","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/2411.02785/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safety alignment of Large Language Models (LLMs) has recently become a critical objective of model developers. In response, a growing body of work has been investigating how safety alignment can be bypassed through various jailbreaking methods, such as adversarial attacks. However, these jailbreak methods can be rather costly or involve a non-trivial amount of creativity and effort, introducing the assumption that malicious users are high-resource or sophisticated. In this paper, we study how simple random augmentations to the input prompt affect safety alignment effectiveness in state-of-the-","authors_text":"Gagandeep Singh, Gaokai Zhang, Hangoo Kang, Jason Vega, Junsheng Huang, Minjia Zhang","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-05T03:51:13Z","title":"Stochastic Monkeys at Play: Random Augmentations Cheaply Break LLM Safety Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.02785","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":"sha256:13fc54d9c83d4b3cbaa9a72498322dc5b25b828ab16d212f2d50deadf7134b1c","target":"record","created_at":"2026-07-05T09:44:54Z","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":"50a7d6c9d4359568beca4d311e7bf34369ab32af446a7cf33de931fb2580571f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-05T03:51:13Z","title_canon_sha256":"9f46a5dbbf4e2638b7de4c39049a77c2ad0f4d6006de1ca80f8fa406c84875cc"},"schema_version":"1.0","source":{"id":"2411.02785","kind":"arxiv","version":2}},"canonical_sha256":"825332e15ad09c2f001246fa7a0f6384a22d11274e48e467046355cc046418a0","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"825332e15ad09c2f001246fa7a0f6384a22d11274e48e467046355cc046418a0","first_computed_at":"2026-07-05T09:44:54.345540Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:44:54.345540Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"w5/2SrMz+SsfvwiMhhFZzSyco8IujHHxg4SVq73eyvSFpLRhFAmIef+RGXTsl4Fq/86/0bPqqoWqXyoB03mFAw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:44:54.346043Z","signed_message":"canonical_sha256_bytes"},"source_id":"2411.02785","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:13fc54d9c83d4b3cbaa9a72498322dc5b25b828ab16d212f2d50deadf7134b1c","sha256:45c0365936fc29464dda897ec8dedb3e60d1a118aad6045de86c30dc18795675"],"state_sha256":"199e5c03036f126390ec7bf378c21a98e3e7131150690ba8ea0a19f2413e7b78"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"93dgZSm7B0lNyK4ZgVtH8sRnbPCcpLbFex5W/4/yHWLkm4/x6TowsGb+s5dFtiMN6Foq0vCOL2gJ2uIhAVg3DQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T13:59:41.510927Z","bundle_sha256":"6861987d041e749d687e31ed88ee11d56952c7b5c05ab24136f07ec467ade37a"}}