{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:NTFZTWS3SDBLNPNJ3FZGUCXZN3","short_pith_number":"pith:NTFZTWS3","canonical_record":{"source":{"id":"2402.03303","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-05T18:58:19Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"be9c96f03d4d1b851a05714d8a297a0457cff67dcb97e8d880168a065e17cbf4","abstract_canon_sha256":"7c3873de23c647fd25b31c3e9e6b8ae852ac316bc03ae06cc5a5904b05cb67f9"},"schema_version":"1.0"},"canonical_sha256":"6ccb99da5b90c2b6bda9d9726a0af96ef74a4c88b756d8c15896c74fabeecb89","source":{"kind":"arxiv","id":"2402.03303","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.03303","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"arxiv_version","alias_value":"2402.03303v1","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.03303","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"pith_short_12","alias_value":"NTFZTWS3SDBL","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"pith_short_16","alias_value":"NTFZTWS3SDBLNPNJ","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"pith_short_8","alias_value":"NTFZTWS3","created_at":"2026-07-05T07:41:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:NTFZTWS3SDBLNPNJ3FZGUCXZN3","target":"record","payload":{"canonical_record":{"source":{"id":"2402.03303","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-05T18:58:19Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"be9c96f03d4d1b851a05714d8a297a0457cff67dcb97e8d880168a065e17cbf4","abstract_canon_sha256":"7c3873de23c647fd25b31c3e9e6b8ae852ac316bc03ae06cc5a5904b05cb67f9"},"schema_version":"1.0"},"canonical_sha256":"6ccb99da5b90c2b6bda9d9726a0af96ef74a4c88b756d8c15896c74fabeecb89","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:41:33.688726Z","signature_b64":"x897oHFeAvzD8igzbOtm+y6zhyCzVjoevUFECI/wSta34HFYaQYy/2V+PQ31deambvOxwdso/qV/YG4FzTGFAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ccb99da5b90c2b6bda9d9726a0af96ef74a4c88b756d8c15896c74fabeecb89","last_reissued_at":"2026-07-05T07:41:33.688145Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:41:33.688145Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.03303","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":"2026-07-05T07:41:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"BsO27DkbCMLHjsPLxpWqY4VLOmprJ5hSRHSSgXkjF1J5jFd0giwn6yacmUJGZEn3lsFPLwo6yVlnZVyW1FRoBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T19:46:55.826410Z"},"content_sha256":"3564ceeac94c81e1a01e8f8d667f59edebe85361a2eb61cd92bb13c838fad128","schema_version":"1.0","event_id":"sha256:3564ceeac94c81e1a01e8f8d667f59edebe85361a2eb61cd92bb13c838fad128"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:NTFZTWS3SDBLNPNJ3FZGUCXZN3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Nevermind: Instruction Override and Moderation in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Edward Kim","submitted_at":"2024-02-05T18:58:19Z","abstract_excerpt":"Given the impressive capabilities of recent Large Language Models (LLMs), we investigate and benchmark the most popular proprietary and different sized open source models on the task of explicit instruction following in conflicting situations, e.g. overrides. These include the ability of the model to override the knowledge within the weights of the model, the ability to override (or moderate) extracted knowledge in the prompt, and lastly the ability to perform a full jailbreak. Experimentation performed suggest several key findings to improve instruction following - larger models perform the b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.03303","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/2402.03303/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":"2026-07-05T07:41:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"AHYsQKNaxQ8/4tEjncK/OQ6we8BCoSCYQcvKfX7x7Y5K1iLbz8l9VD5tFf7M9Hz+5f+5XInvCrdS7zz+19qWDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T19:46:55.827684Z"},"content_sha256":"2f050a7235c06db367fac3b4d6cde9d41e851a908e092c02ac1f8c7b27981e36","schema_version":"1.0","event_id":"sha256:2f050a7235c06db367fac3b4d6cde9d41e851a908e092c02ac1f8c7b27981e36"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/NTFZTWS3SDBLNPNJ3FZGUCXZN3/bundle.json","state_url":"https://pith.science/pith/NTFZTWS3SDBLNPNJ3FZGUCXZN3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/NTFZTWS3SDBLNPNJ3FZGUCXZN3/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":"2026-08-09T19:46:55Z","links":{"resolver":"https://pith.science/pith/NTFZTWS3SDBLNPNJ3FZGUCXZN3","bundle":"https://pith.science/pith/NTFZTWS3SDBLNPNJ3FZGUCXZN3/bundle.json","state":"https://pith.science/pith/NTFZTWS3SDBLNPNJ3FZGUCXZN3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/NTFZTWS3SDBLNPNJ3FZGUCXZN3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:NTFZTWS3SDBLNPNJ3FZGUCXZN3","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":"7c3873de23c647fd25b31c3e9e6b8ae852ac316bc03ae06cc5a5904b05cb67f9","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-05T18:58:19Z","title_canon_sha256":"be9c96f03d4d1b851a05714d8a297a0457cff67dcb97e8d880168a065e17cbf4"},"schema_version":"1.0","source":{"id":"2402.03303","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.03303","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"arxiv_version","alias_value":"2402.03303v1","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.03303","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"pith_short_12","alias_value":"NTFZTWS3SDBL","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"pith_short_16","alias_value":"NTFZTWS3SDBLNPNJ","created_at":"2026-07-05T07:41:33Z"},{"alias_kind":"pith_short_8","alias_value":"NTFZTWS3","created_at":"2026-07-05T07:41:33Z"}],"graph_snapshots":[{"event_id":"sha256:2f050a7235c06db367fac3b4d6cde9d41e851a908e092c02ac1f8c7b27981e36","target":"graph","created_at":"2026-07-05T07:41:33Z","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/2402.03303/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Given the impressive capabilities of recent Large Language Models (LLMs), we investigate and benchmark the most popular proprietary and different sized open source models on the task of explicit instruction following in conflicting situations, e.g. overrides. These include the ability of the model to override the knowledge within the weights of the model, the ability to override (or moderate) extracted knowledge in the prompt, and lastly the ability to perform a full jailbreak. Experimentation performed suggest several key findings to improve instruction following - larger models perform the b","authors_text":"Edward Kim","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-05T18:58:19Z","title":"Nevermind: Instruction Override and Moderation in Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.03303","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":"sha256:3564ceeac94c81e1a01e8f8d667f59edebe85361a2eb61cd92bb13c838fad128","target":"record","created_at":"2026-07-05T07:41:33Z","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":"7c3873de23c647fd25b31c3e9e6b8ae852ac316bc03ae06cc5a5904b05cb67f9","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-05T18:58:19Z","title_canon_sha256":"be9c96f03d4d1b851a05714d8a297a0457cff67dcb97e8d880168a065e17cbf4"},"schema_version":"1.0","source":{"id":"2402.03303","kind":"arxiv","version":1}},"canonical_sha256":"6ccb99da5b90c2b6bda9d9726a0af96ef74a4c88b756d8c15896c74fabeecb89","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6ccb99da5b90c2b6bda9d9726a0af96ef74a4c88b756d8c15896c74fabeecb89","first_computed_at":"2026-07-05T07:41:33.688145Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:41:33.688145Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"x897oHFeAvzD8igzbOtm+y6zhyCzVjoevUFECI/wSta34HFYaQYy/2V+PQ31deambvOxwdso/qV/YG4FzTGFAA==","signature_status":"signed_v1","signed_at":"2026-07-05T07:41:33.688726Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.03303","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3564ceeac94c81e1a01e8f8d667f59edebe85361a2eb61cd92bb13c838fad128","sha256:2f050a7235c06db367fac3b4d6cde9d41e851a908e092c02ac1f8c7b27981e36"],"state_sha256":"10cf9ee0e393225be594184f4843e7c4ac657006b1870c404994e84b181af838"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ur/UrWu3Fm/+agKtFmn48Telqu1MStQXHvmB5Grbstw8/GE1Z7P3f0jmimRJ2wLUIY5T0l6AY6/isiXsbrJ2CQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T19:46:55.837341Z","bundle_sha256":"9cef84245c241e456baac7d8fa48336cde1f2bdad84902695d42906d0fb35f39"}}