{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:W424NIQGYCGE2PDPLXMSVTCSDS","short_pith_number":"pith:W424NIQG","canonical_record":{"source":{"id":"2410.21965","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T11:47:01Z","cross_cats_sorted":[],"title_canon_sha256":"0d3c72e13435c0d19acb2202c12fe6f7c5187968e1ed1ef72e090b9a013ffea2","abstract_canon_sha256":"0a5f34124c09b673686bd2a4e173ecc7b2a4ffd18285608d60b50890bb643488"},"schema_version":"1.0"},"canonical_sha256":"b735c6a206c08c4d3c6f5dd92acc521c988bc6f31eb6f520dc3bdcb17184b057","source":{"kind":"arxiv","id":"2410.21965","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.21965","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"arxiv_version","alias_value":"2410.21965v1","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.21965","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"pith_short_12","alias_value":"W424NIQGYCGE","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"pith_short_16","alias_value":"W424NIQGYCGE2PDP","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"pith_short_8","alias_value":"W424NIQG","created_at":"2026-07-05T09:28:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:W424NIQGYCGE2PDPLXMSVTCSDS","target":"record","payload":{"canonical_record":{"source":{"id":"2410.21965","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T11:47:01Z","cross_cats_sorted":[],"title_canon_sha256":"0d3c72e13435c0d19acb2202c12fe6f7c5187968e1ed1ef72e090b9a013ffea2","abstract_canon_sha256":"0a5f34124c09b673686bd2a4e173ecc7b2a4ffd18285608d60b50890bb643488"},"schema_version":"1.0"},"canonical_sha256":"b735c6a206c08c4d3c6f5dd92acc521c988bc6f31eb6f520dc3bdcb17184b057","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:28:02.263608Z","signature_b64":"cLxNBlQexEfDlF+h0NQMbuOqcxvd6yvkRptq4gXhmzN4/BIhGalHSVbCP2Ml2xjUhiqBTbdPYiEgPUuarYYlCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b735c6a206c08c4d3c6f5dd92acc521c988bc6f31eb6f520dc3bdcb17184b057","last_reissued_at":"2026-07-05T09:28:02.263019Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:28:02.263019Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.21965","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":"2026-07-05T09:28:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qT90LAl85svbPoPWY/C/MKvUHiYje/5+uP426A/JcwHo7lTrlus3p6Pra7vXGhPTXO4nQxW2PSYZ80EpOtMJDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T17:19:09.634972Z"},"content_sha256":"498c5ae3204c4398bc9be050177ebf2eba3231492dd5edccf01d1482a3379dda","schema_version":"1.0","event_id":"sha256:498c5ae3204c4398bc9be050177ebf2eba3231492dd5edccf01d1482a3379dda"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:W424NIQGYCGE2PDPLXMSVTCSDS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SG-Bench: Evaluating LLM Safety Generalization Across Diverse Tasks and Prompt Types","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Shikun Zhang, Wei Ye, Yutao Mou","submitted_at":"2024-10-29T11:47:01Z","abstract_excerpt":"Ensuring the safety of large language model (LLM) applications is essential for developing trustworthy artificial intelligence. Current LLM safety benchmarks have two limitations. First, they focus solely on either discriminative or generative evaluation paradigms while ignoring their interconnection. Second, they rely on standardized inputs, overlooking the effects of widespread prompting techniques, such as system prompts, few-shot demonstrations, and chain-of-thought prompting. To overcome these issues, we developed SG-Bench, a novel benchmark to assess the generalization of LLM safety acro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.21965","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/2410.21965/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":"2026-07-05T09:28:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oR8/hjQB009L9lCVGuOVW8IvVKNnMllPgh0WH2g3JKUVTeAwMCZu5ORoNlj8I0C1Qn7QTgiMI/0b7D4y+ogiBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T17:19:09.635349Z"},"content_sha256":"7f1f0566288355c94f25d4b59b838d26a5e12a6435b639a51a6133e8575672c7","schema_version":"1.0","event_id":"sha256:7f1f0566288355c94f25d4b59b838d26a5e12a6435b639a51a6133e8575672c7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/W424NIQGYCGE2PDPLXMSVTCSDS/bundle.json","state_url":"https://pith.science/pith/W424NIQGYCGE2PDPLXMSVTCSDS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/W424NIQGYCGE2PDPLXMSVTCSDS/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":"2026-07-06T17:19:09Z","links":{"resolver":"https://pith.science/pith/W424NIQGYCGE2PDPLXMSVTCSDS","bundle":"https://pith.science/pith/W424NIQGYCGE2PDPLXMSVTCSDS/bundle.json","state":"https://pith.science/pith/W424NIQGYCGE2PDPLXMSVTCSDS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/W424NIQGYCGE2PDPLXMSVTCSDS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:W424NIQGYCGE2PDPLXMSVTCSDS","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":"0a5f34124c09b673686bd2a4e173ecc7b2a4ffd18285608d60b50890bb643488","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T11:47:01Z","title_canon_sha256":"0d3c72e13435c0d19acb2202c12fe6f7c5187968e1ed1ef72e090b9a013ffea2"},"schema_version":"1.0","source":{"id":"2410.21965","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.21965","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"arxiv_version","alias_value":"2410.21965v1","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.21965","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"pith_short_12","alias_value":"W424NIQGYCGE","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"pith_short_16","alias_value":"W424NIQGYCGE2PDP","created_at":"2026-07-05T09:28:02Z"},{"alias_kind":"pith_short_8","alias_value":"W424NIQG","created_at":"2026-07-05T09:28:02Z"}],"graph_snapshots":[{"event_id":"sha256:7f1f0566288355c94f25d4b59b838d26a5e12a6435b639a51a6133e8575672c7","target":"graph","created_at":"2026-07-05T09:28:02Z","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/2410.21965/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Ensuring the safety of large language model (LLM) applications is essential for developing trustworthy artificial intelligence. Current LLM safety benchmarks have two limitations. First, they focus solely on either discriminative or generative evaluation paradigms while ignoring their interconnection. Second, they rely on standardized inputs, overlooking the effects of widespread prompting techniques, such as system prompts, few-shot demonstrations, and chain-of-thought prompting. To overcome these issues, we developed SG-Bench, a novel benchmark to assess the generalization of LLM safety acro","authors_text":"Shikun Zhang, Wei Ye, Yutao Mou","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T11:47:01Z","title":"SG-Bench: Evaluating LLM Safety Generalization Across Diverse Tasks and Prompt Types"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.21965","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":"sha256:498c5ae3204c4398bc9be050177ebf2eba3231492dd5edccf01d1482a3379dda","target":"record","created_at":"2026-07-05T09:28:02Z","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":"0a5f34124c09b673686bd2a4e173ecc7b2a4ffd18285608d60b50890bb643488","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-29T11:47:01Z","title_canon_sha256":"0d3c72e13435c0d19acb2202c12fe6f7c5187968e1ed1ef72e090b9a013ffea2"},"schema_version":"1.0","source":{"id":"2410.21965","kind":"arxiv","version":1}},"canonical_sha256":"b735c6a206c08c4d3c6f5dd92acc521c988bc6f31eb6f520dc3bdcb17184b057","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b735c6a206c08c4d3c6f5dd92acc521c988bc6f31eb6f520dc3bdcb17184b057","first_computed_at":"2026-07-05T09:28:02.263019Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:28:02.263019Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"cLxNBlQexEfDlF+h0NQMbuOqcxvd6yvkRptq4gXhmzN4/BIhGalHSVbCP2Ml2xjUhiqBTbdPYiEgPUuarYYlCA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:28:02.263608Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.21965","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:498c5ae3204c4398bc9be050177ebf2eba3231492dd5edccf01d1482a3379dda","sha256:7f1f0566288355c94f25d4b59b838d26a5e12a6435b639a51a6133e8575672c7"],"state_sha256":"cf88419fa9350175c5085708912a7741f021a1815a11abd78b3397556dbb0e3b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"b95eOZUbPpfXqm7CAAkFtkd9LmXoHvIgbMjhepLwuR9a774+s1+hnUBxlQtAjRVRaYhmat5a3XbIZ1ldSIHlBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-06T17:19:09.637263Z","bundle_sha256":"d2d0cbafd21a68ac92a69a8c463bee23ac745c5edb055f9051e833ef3ace9650"}}