{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:N4FIMP7NOLK4NYDUP2PNVLOV5B","short_pith_number":"pith:N4FIMP7N","canonical_record":{"source":{"id":"2401.00287","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2023-12-30T17:37:06Z","cross_cats_sorted":[],"title_canon_sha256":"d85cd49660a99a3868f07161f662c65be295ea3da597b8df78f248151e24893e","abstract_canon_sha256":"01191c03fcca11a5f9df57c0e2b8639234165bd44e04d82f67c5e417e25491b9"},"schema_version":"1.0"},"canonical_sha256":"6f0a863fed72d5c6e0747e9edaadd5e85bfaf3376758fc53ddfefc579192dd28","source":{"kind":"arxiv","id":"2401.00287","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.00287","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"arxiv_version","alias_value":"2401.00287v1","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.00287","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"pith_short_12","alias_value":"N4FIMP7NOLK4","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"pith_short_16","alias_value":"N4FIMP7NOLK4NYDU","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"pith_short_8","alias_value":"N4FIMP7N","created_at":"2026-07-05T07:29:15Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:N4FIMP7NOLK4NYDUP2PNVLOV5B","target":"record","payload":{"canonical_record":{"source":{"id":"2401.00287","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2023-12-30T17:37:06Z","cross_cats_sorted":[],"title_canon_sha256":"d85cd49660a99a3868f07161f662c65be295ea3da597b8df78f248151e24893e","abstract_canon_sha256":"01191c03fcca11a5f9df57c0e2b8639234165bd44e04d82f67c5e417e25491b9"},"schema_version":"1.0"},"canonical_sha256":"6f0a863fed72d5c6e0747e9edaadd5e85bfaf3376758fc53ddfefc579192dd28","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:29:15.170047Z","signature_b64":"1h3/S8basoMhieeu/r34quySnWRD0XTdnbS1yMXg3P2zHo7XgNA9jHSd33Ge7bpbOcVHWu1c18sIQQtSynB4CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6f0a863fed72d5c6e0747e9edaadd5e85bfaf3376758fc53ddfefc579192dd28","last_reissued_at":"2026-07-05T07:29:15.169588Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:29:15.169588Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2401.00287","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":"2026-07-05T07:29:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1R5ZEKBpw7tsWIRdbY/5YwA/BBYNFm4WeGKRRBgH2XudTl7f2duSpXQrUvSyo5mW/y97Z+IuRlahYPV80JB9Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-16T02:06:36.303672Z"},"content_sha256":"8b4b28520c5e6af6331f914a10f2334a82fe14ec9357f9fd88088bbf47403bcc","schema_version":"1.0","event_id":"sha256:8b4b28520c5e6af6331f914a10f2334a82fe14ec9357f9fd88088bbf47403bcc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:N4FIMP7NOLK4NYDUP2PNVLOV5B","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"The Art of Defending: A Systematic Evaluation and Analysis of LLM Defense Strategies on Safety and Over-Defensiveness","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Agastya Seth, Chitta Baral, Neeraj Varshney, Pavel Dolin","submitted_at":"2023-12-30T17:37:06Z","abstract_excerpt":"As Large Language Models (LLMs) play an increasingly pivotal role in natural language processing applications, their safety concerns become critical areas of NLP research. This paper presents Safety and Over-Defensiveness Evaluation (SODE) benchmark: a collection of diverse safe and unsafe prompts with carefully designed evaluation methods that facilitate systematic evaluation, comparison, and analysis over 'safety' and 'over-defensiveness.' With SODE, we study a variety of LLM defense strategies over multiple state-of-the-art LLMs, which reveals several interesting and important findings, suc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.00287","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/2401.00287/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":"2026-07-05T07:29:15Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8TQCYNDxLazIqfsCNZ5yOmddUBwN9JRt/+d3NdqX/kyp1dY+YPYsIjhY55fCeflz5B8HXZaqP5RLnSdmy64iDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-16T02:06:36.304059Z"},"content_sha256":"e718e807e0b17a445761d1de3fb33925ba6c1d6cd159fe8986990c59e8bf0fea","schema_version":"1.0","event_id":"sha256:e718e807e0b17a445761d1de3fb33925ba6c1d6cd159fe8986990c59e8bf0fea"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/N4FIMP7NOLK4NYDUP2PNVLOV5B/bundle.json","state_url":"https://pith.science/pith/N4FIMP7NOLK4NYDUP2PNVLOV5B/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/N4FIMP7NOLK4NYDUP2PNVLOV5B/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":"2026-07-16T02:06:36Z","links":{"resolver":"https://pith.science/pith/N4FIMP7NOLK4NYDUP2PNVLOV5B","bundle":"https://pith.science/pith/N4FIMP7NOLK4NYDUP2PNVLOV5B/bundle.json","state":"https://pith.science/pith/N4FIMP7NOLK4NYDUP2PNVLOV5B/state.json","well_known_bundle":"https://pith.science/.well-known/pith/N4FIMP7NOLK4NYDUP2PNVLOV5B/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:N4FIMP7NOLK4NYDUP2PNVLOV5B","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":"01191c03fcca11a5f9df57c0e2b8639234165bd44e04d82f67c5e417e25491b9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2023-12-30T17:37:06Z","title_canon_sha256":"d85cd49660a99a3868f07161f662c65be295ea3da597b8df78f248151e24893e"},"schema_version":"1.0","source":{"id":"2401.00287","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.00287","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"arxiv_version","alias_value":"2401.00287v1","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.00287","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"pith_short_12","alias_value":"N4FIMP7NOLK4","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"pith_short_16","alias_value":"N4FIMP7NOLK4NYDU","created_at":"2026-07-05T07:29:15Z"},{"alias_kind":"pith_short_8","alias_value":"N4FIMP7N","created_at":"2026-07-05T07:29:15Z"}],"graph_snapshots":[{"event_id":"sha256:e718e807e0b17a445761d1de3fb33925ba6c1d6cd159fe8986990c59e8bf0fea","target":"graph","created_at":"2026-07-05T07:29:15Z","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/2401.00287/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As Large Language Models (LLMs) play an increasingly pivotal role in natural language processing applications, their safety concerns become critical areas of NLP research. This paper presents Safety and Over-Defensiveness Evaluation (SODE) benchmark: a collection of diverse safe and unsafe prompts with carefully designed evaluation methods that facilitate systematic evaluation, comparison, and analysis over 'safety' and 'over-defensiveness.' With SODE, we study a variety of LLM defense strategies over multiple state-of-the-art LLMs, which reveals several interesting and important findings, suc","authors_text":"Agastya Seth, Chitta Baral, Neeraj Varshney, Pavel Dolin","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2023-12-30T17:37:06Z","title":"The Art of Defending: A Systematic Evaluation and Analysis of LLM Defense Strategies on Safety and Over-Defensiveness"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.00287","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":"sha256:8b4b28520c5e6af6331f914a10f2334a82fe14ec9357f9fd88088bbf47403bcc","target":"record","created_at":"2026-07-05T07:29:15Z","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":"01191c03fcca11a5f9df57c0e2b8639234165bd44e04d82f67c5e417e25491b9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2023-12-30T17:37:06Z","title_canon_sha256":"d85cd49660a99a3868f07161f662c65be295ea3da597b8df78f248151e24893e"},"schema_version":"1.0","source":{"id":"2401.00287","kind":"arxiv","version":1}},"canonical_sha256":"6f0a863fed72d5c6e0747e9edaadd5e85bfaf3376758fc53ddfefc579192dd28","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6f0a863fed72d5c6e0747e9edaadd5e85bfaf3376758fc53ddfefc579192dd28","first_computed_at":"2026-07-05T07:29:15.169588Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:29:15.169588Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"1h3/S8basoMhieeu/r34quySnWRD0XTdnbS1yMXg3P2zHo7XgNA9jHSd33Ge7bpbOcVHWu1c18sIQQtSynB4CQ==","signature_status":"signed_v1","signed_at":"2026-07-05T07:29:15.170047Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.00287","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8b4b28520c5e6af6331f914a10f2334a82fe14ec9357f9fd88088bbf47403bcc","sha256:e718e807e0b17a445761d1de3fb33925ba6c1d6cd159fe8986990c59e8bf0fea"],"state_sha256":"d99ebaa0f124840df8e317646938e0a47cb4aaf6a28b25fa9b961a0bcd2aaff7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nBc1dzUqOMsRQX/iCW3vA56bgkcfuXhZN/Nkw7A/OwFkUI0u6mSaQKt2bjiQwy+haE7BnBP7ioEGjH5pF+STDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-16T02:06:36.306276Z","bundle_sha256":"1bd3f0f95caa31adf98f5ea87b17b1b33bfe7c404676ec6be2039cb41ce2e07a"}}