{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:IMVXQBWLNI3LNWMO3IV72WN4FT","short_pith_number":"pith:IMVXQBWL","canonical_record":{"source":{"id":"2402.05723","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-08T14:54:17Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"08491d53b9e3b030516c8525da691889e545142be8ce1fbc6da7e74010f7c99a","abstract_canon_sha256":"29c2bd8d57310620577301cabd1d38701554bc9ec7c129bef144aabebc415f82"},"schema_version":"1.0"},"canonical_sha256":"432b7806cb6a36b6d98eda2bfd59bc2ccdab901fcad7aa3787f88e9b2855f465","source":{"kind":"arxiv","id":"2402.05723","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.05723","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"arxiv_version","alias_value":"2402.05723v1","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.05723","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"pith_short_12","alias_value":"IMVXQBWLNI3L","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"pith_short_16","alias_value":"IMVXQBWLNI3LNWMO","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"pith_short_8","alias_value":"IMVXQBWL","created_at":"2026-07-05T07:43:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:IMVXQBWLNI3LNWMO3IV72WN4FT","target":"record","payload":{"canonical_record":{"source":{"id":"2402.05723","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-08T14:54:17Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"08491d53b9e3b030516c8525da691889e545142be8ce1fbc6da7e74010f7c99a","abstract_canon_sha256":"29c2bd8d57310620577301cabd1d38701554bc9ec7c129bef144aabebc415f82"},"schema_version":"1.0"},"canonical_sha256":"432b7806cb6a36b6d98eda2bfd59bc2ccdab901fcad7aa3787f88e9b2855f465","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:43:02.526471Z","signature_b64":"x4qrpRwHzS1HOB8miSWUuEP9/HZGojE7UH4+SMxq8STLZZY79sFIRc8MYx7VRIvkMdBHQbVwY7GlfCLoSYghBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"432b7806cb6a36b6d98eda2bfd59bc2ccdab901fcad7aa3787f88e9b2855f465","last_reissued_at":"2026-07-05T07:43:02.526011Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:43:02.526011Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.05723","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":"2026-07-05T07:43:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PXRzh1sTp8guPFoECkurtAm9P5DzvgIlZyYVuKm2aOnzbvbjxbr8/xBS6ZxW3eIS8xG4YIDzb9+zpUSYZkL4Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T18:03:11.689270Z"},"content_sha256":"90b43376121cdfc0ea50fb3485a59e949d9f11eaec6450c9ac214061b8f43921","schema_version":"1.0","event_id":"sha256:90b43376121cdfc0ea50fb3485a59e949d9f11eaec6450c9ac214061b8f43921"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:IMVXQBWLNI3LNWMO3IV72WN4FT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"In-Context Learning Can Re-learn Forbidden Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"David Dobre, Dhanya Sridhar, Gauthier Gidel, Jian Tang, Sophie Xhonneux","submitted_at":"2024-02-08T14:54:17Z","abstract_excerpt":"Despite significant investment into safety training, large language models (LLMs) deployed in the real world still suffer from numerous vulnerabilities. One perspective on LLM safety training is that it algorithmically forbids the model from answering toxic or harmful queries. To assess the effectiveness of safety training, in this work, we study forbidden tasks, i.e., tasks the model is designed to refuse to answer. Specifically, we investigate whether in-context learning (ICL) can be used to re-learn forbidden tasks despite the explicit fine-tuning of the model to refuse them. We first exami"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.05723","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/2402.05723/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":"2026-07-05T07:43:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"K7N+4yvr2SJvwM43bEcbGILGZNdcEh/SLoXrtGvA6g3Xp7DORrRZ8F4C3gfiOYYZUtNYgYWW4jmupsj5Y0YHBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T18:03:11.689757Z"},"content_sha256":"37c6de780356f75266174c9e1ca55514f960047385372d4d90ab93d89b535d0c","schema_version":"1.0","event_id":"sha256:37c6de780356f75266174c9e1ca55514f960047385372d4d90ab93d89b535d0c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IMVXQBWLNI3LNWMO3IV72WN4FT/bundle.json","state_url":"https://pith.science/pith/IMVXQBWLNI3LNWMO3IV72WN4FT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IMVXQBWLNI3LNWMO3IV72WN4FT/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":"2026-08-08T18:03:11Z","links":{"resolver":"https://pith.science/pith/IMVXQBWLNI3LNWMO3IV72WN4FT","bundle":"https://pith.science/pith/IMVXQBWLNI3LNWMO3IV72WN4FT/bundle.json","state":"https://pith.science/pith/IMVXQBWLNI3LNWMO3IV72WN4FT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IMVXQBWLNI3LNWMO3IV72WN4FT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:IMVXQBWLNI3LNWMO3IV72WN4FT","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":"29c2bd8d57310620577301cabd1d38701554bc9ec7c129bef144aabebc415f82","cross_cats_sorted":["cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-08T14:54:17Z","title_canon_sha256":"08491d53b9e3b030516c8525da691889e545142be8ce1fbc6da7e74010f7c99a"},"schema_version":"1.0","source":{"id":"2402.05723","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.05723","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"arxiv_version","alias_value":"2402.05723v1","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.05723","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"pith_short_12","alias_value":"IMVXQBWLNI3L","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"pith_short_16","alias_value":"IMVXQBWLNI3LNWMO","created_at":"2026-07-05T07:43:02Z"},{"alias_kind":"pith_short_8","alias_value":"IMVXQBWL","created_at":"2026-07-05T07:43:02Z"}],"graph_snapshots":[{"event_id":"sha256:37c6de780356f75266174c9e1ca55514f960047385372d4d90ab93d89b535d0c","target":"graph","created_at":"2026-07-05T07:43:02Z","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/2402.05723/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite significant investment into safety training, large language models (LLMs) deployed in the real world still suffer from numerous vulnerabilities. One perspective on LLM safety training is that it algorithmically forbids the model from answering toxic or harmful queries. To assess the effectiveness of safety training, in this work, we study forbidden tasks, i.e., tasks the model is designed to refuse to answer. Specifically, we investigate whether in-context learning (ICL) can be used to re-learn forbidden tasks despite the explicit fine-tuning of the model to refuse them. We first exami","authors_text":"David Dobre, Dhanya Sridhar, Gauthier Gidel, Jian Tang, Sophie Xhonneux","cross_cats":["cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-08T14:54:17Z","title":"In-Context Learning Can Re-learn Forbidden Tasks"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.05723","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":"sha256:90b43376121cdfc0ea50fb3485a59e949d9f11eaec6450c9ac214061b8f43921","target":"record","created_at":"2026-07-05T07:43:02Z","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":"29c2bd8d57310620577301cabd1d38701554bc9ec7c129bef144aabebc415f82","cross_cats_sorted":["cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-08T14:54:17Z","title_canon_sha256":"08491d53b9e3b030516c8525da691889e545142be8ce1fbc6da7e74010f7c99a"},"schema_version":"1.0","source":{"id":"2402.05723","kind":"arxiv","version":1}},"canonical_sha256":"432b7806cb6a36b6d98eda2bfd59bc2ccdab901fcad7aa3787f88e9b2855f465","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"432b7806cb6a36b6d98eda2bfd59bc2ccdab901fcad7aa3787f88e9b2855f465","first_computed_at":"2026-07-05T07:43:02.526011Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:43:02.526011Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"x4qrpRwHzS1HOB8miSWUuEP9/HZGojE7UH4+SMxq8STLZZY79sFIRc8MYx7VRIvkMdBHQbVwY7GlfCLoSYghBw==","signature_status":"signed_v1","signed_at":"2026-07-05T07:43:02.526471Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.05723","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:90b43376121cdfc0ea50fb3485a59e949d9f11eaec6450c9ac214061b8f43921","sha256:37c6de780356f75266174c9e1ca55514f960047385372d4d90ab93d89b535d0c"],"state_sha256":"7cbd1f13bbf7c55fcb80b3173e4f5ef5c0d7130f97558a905a2c411b36484233"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2DWP8gyrfHxF5Gxi7stu0D4WQpdAfcDfDIG51O7fZfYatDMUQrst7WBBmEHiKVtzEldVyZzGLO/VZ/Bx9XAhDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T18:03:11.693208Z","bundle_sha256":"4785c0c1490bfee62642dfc0c18137930aed72b077bdf5dfec69097966f41501"}}