{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:JHMHFF5RP7DN7ANYNXHN2RAW4Q","short_pith_number":"pith:JHMHFF5R","canonical_record":{"source":{"id":"2310.02949","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-04T16:39:31Z","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"title_canon_sha256":"32267b5d46f9b61a065b3c5bb49538a40c7bf884871d6dd1ccf1ce3364f5e797","abstract_canon_sha256":"93581bc90d7e03c1a54b0c6e5a241a872150228feae6648637e1fde130501c3b"},"schema_version":"1.0"},"canonical_sha256":"49d87297b17fc6df81b86dcedd4416e4398953f5afcc75aed3a3c20f25f28710","source":{"kind":"arxiv","id":"2310.02949","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.02949","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"arxiv_version","alias_value":"2310.02949v1","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.02949","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"pith_short_12","alias_value":"JHMHFF5RP7DN","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"pith_short_16","alias_value":"JHMHFF5RP7DN7ANY","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"pith_short_8","alias_value":"JHMHFF5R","created_at":"2026-07-05T06:57:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:JHMHFF5RP7DN7ANYNXHN2RAW4Q","target":"record","payload":{"canonical_record":{"source":{"id":"2310.02949","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-04T16:39:31Z","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"title_canon_sha256":"32267b5d46f9b61a065b3c5bb49538a40c7bf884871d6dd1ccf1ce3364f5e797","abstract_canon_sha256":"93581bc90d7e03c1a54b0c6e5a241a872150228feae6648637e1fde130501c3b"},"schema_version":"1.0"},"canonical_sha256":"49d87297b17fc6df81b86dcedd4416e4398953f5afcc75aed3a3c20f25f28710","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:57:15.818245Z","signature_b64":"EirjBWEvCa9rqP6R0KVOGzLXVyGF6M+irrgs7QjiRhXEROARA39oB8mrdEsZxuhKg4IX91GGYrjwyUDinE/xAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"49d87297b17fc6df81b86dcedd4416e4398953f5afcc75aed3a3c20f25f28710","last_reissued_at":"2026-07-05T06:57:15.817757Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:57:15.817757Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2310.02949","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":"2026-07-05T06:57:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+MtbzFFdmsAm8F9+ZSAY8iYHhu9s7QijDngSxyWVtjfMe0lNvlbi8hyF/uq8Ml1XF5nzVshrC0ynFgShyNz7AA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T18:08:55.168263Z"},"content_sha256":"d5b54e9a9fa435608cfeb20c4e2108f2a831bc928670aed990786449f6595c70","schema_version":"1.0","event_id":"sha256:d5b54e9a9fa435608cfeb20c4e2108f2a831bc928670aed990786449f6595c70"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:JHMHFF5RP7DN7ANYNXHN2RAW4Q","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Dahua Lin, Linda Petzold, Qi Zhang, William Yang Wang, Xianjun Yang, Xiao Wang, Xun Zhao","submitted_at":"2023-10-04T16:39:31Z","abstract_excerpt":"Warning: This paper contains examples of harmful language, and reader discretion is recommended. The increasing open release of powerful large language models (LLMs) has facilitated the development of downstream applications by reducing the essential cost of data annotation and computation. To ensure AI safety, extensive safety-alignment measures have been conducted to armor these models against malicious use (primarily hard prompt attack). However, beneath the seemingly resilient facade of the armor, there might lurk a shadow. By simply tuning on 100 malicious examples with 1 GPU hour, these "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.02949","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/2310.02949/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":"2026-07-05T06:57:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Ar/Xw4/nZBnpGjuJUgPvQ6qjSOT/e0wmyiNiadz5ay4u8Avw710GycHd9VLhhb1Ga/dmD+L01/yGM/NaHqpqBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T18:08:55.168793Z"},"content_sha256":"45e49df5ee62bb766c7dd241f37d556c0d144d27610a42949fc24f3da35d94cf","schema_version":"1.0","event_id":"sha256:45e49df5ee62bb766c7dd241f37d556c0d144d27610a42949fc24f3da35d94cf"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JHMHFF5RP7DN7ANYNXHN2RAW4Q/bundle.json","state_url":"https://pith.science/pith/JHMHFF5RP7DN7ANYNXHN2RAW4Q/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JHMHFF5RP7DN7ANYNXHN2RAW4Q/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":"2026-08-07T18:08:55Z","links":{"resolver":"https://pith.science/pith/JHMHFF5RP7DN7ANYNXHN2RAW4Q","bundle":"https://pith.science/pith/JHMHFF5RP7DN7ANYNXHN2RAW4Q/bundle.json","state":"https://pith.science/pith/JHMHFF5RP7DN7ANYNXHN2RAW4Q/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JHMHFF5RP7DN7ANYNXHN2RAW4Q/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:JHMHFF5RP7DN7ANYNXHN2RAW4Q","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":"93581bc90d7e03c1a54b0c6e5a241a872150228feae6648637e1fde130501c3b","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-04T16:39:31Z","title_canon_sha256":"32267b5d46f9b61a065b3c5bb49538a40c7bf884871d6dd1ccf1ce3364f5e797"},"schema_version":"1.0","source":{"id":"2310.02949","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.02949","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"arxiv_version","alias_value":"2310.02949v1","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.02949","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"pith_short_12","alias_value":"JHMHFF5RP7DN","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"pith_short_16","alias_value":"JHMHFF5RP7DN7ANY","created_at":"2026-07-05T06:57:15Z"},{"alias_kind":"pith_short_8","alias_value":"JHMHFF5R","created_at":"2026-07-05T06:57:15Z"}],"graph_snapshots":[{"event_id":"sha256:45e49df5ee62bb766c7dd241f37d556c0d144d27610a42949fc24f3da35d94cf","target":"graph","created_at":"2026-07-05T06:57:15Z","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/2310.02949/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Warning: This paper contains examples of harmful language, and reader discretion is recommended. The increasing open release of powerful large language models (LLMs) has facilitated the development of downstream applications by reducing the essential cost of data annotation and computation. To ensure AI safety, extensive safety-alignment measures have been conducted to armor these models against malicious use (primarily hard prompt attack). However, beneath the seemingly resilient facade of the armor, there might lurk a shadow. By simply tuning on 100 malicious examples with 1 GPU hour, these ","authors_text":"Dahua Lin, Linda Petzold, Qi Zhang, William Yang Wang, Xianjun Yang, Xiao Wang, Xun Zhao","cross_cats":["cs.AI","cs.CR","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-04T16:39:31Z","title":"Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.02949","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:d5b54e9a9fa435608cfeb20c4e2108f2a831bc928670aed990786449f6595c70","target":"record","created_at":"2026-07-05T06:57:15Z","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":"93581bc90d7e03c1a54b0c6e5a241a872150228feae6648637e1fde130501c3b","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-04T16:39:31Z","title_canon_sha256":"32267b5d46f9b61a065b3c5bb49538a40c7bf884871d6dd1ccf1ce3364f5e797"},"schema_version":"1.0","source":{"id":"2310.02949","kind":"arxiv","version":1}},"canonical_sha256":"49d87297b17fc6df81b86dcedd4416e4398953f5afcc75aed3a3c20f25f28710","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"49d87297b17fc6df81b86dcedd4416e4398953f5afcc75aed3a3c20f25f28710","first_computed_at":"2026-07-05T06:57:15.817757Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T06:57:15.817757Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"EirjBWEvCa9rqP6R0KVOGzLXVyGF6M+irrgs7QjiRhXEROARA39oB8mrdEsZxuhKg4IX91GGYrjwyUDinE/xAw==","signature_status":"signed_v1","signed_at":"2026-07-05T06:57:15.818245Z","signed_message":"canonical_sha256_bytes"},"source_id":"2310.02949","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d5b54e9a9fa435608cfeb20c4e2108f2a831bc928670aed990786449f6595c70","sha256:45e49df5ee62bb766c7dd241f37d556c0d144d27610a42949fc24f3da35d94cf"],"state_sha256":"a655cd9241f60a8a331b1fc4781be12df7088f9865f674ebb02fe8742b2760a0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LbrgC9plmG5rMYiFyyj1tilUbCDOmvfEkGUtqd4YjkfqiUnVAQ23i6fKyYZqQZqf1upXK1eo7W89PIc/gx3pBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T18:08:55.173892Z","bundle_sha256":"b0aa7144c73e4f63d0bc3d9db2168f6785a0b376f2a16ad290b0783c0645dd1c"}}