{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7QGOM3GZKRC3LL4RPA6UNE63AK","short_pith_number":"pith:7QGOM3GZ","schema_version":"1.0","canonical_sha256":"fc0ce66cd95445b5af91783d4693db02b962913d5ca7a58fac4a2dd6a872f9c6","source":{"kind":"arxiv","id":"2502.16971","version":1},"attestation_state":"computed","paper":{"title":"LongSafety: Evaluating Long-Context Safety of Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Cunxiang Wang, Hongning Wang, Jiale Cheng, Jie Tang, Minlie Huang, Shiyao Cui, Xiaotao Gu, Yida Lu, Yuxiao Dong, Zhexin Zhang","submitted_at":"2025-02-24T08:54:39Z","abstract_excerpt":"As Large Language Models (LLMs) continue to advance in understanding and generating long sequences, new safety concerns have been introduced through the long context. However, the safety of LLMs in long-context tasks remains under-explored, leaving a significant gap in both evaluation and improvement of their safety. To address this, we introduce LongSafety, the first comprehensive benchmark specifically designed to evaluate LLM safety in open-ended long-context tasks. LongSafety encompasses 7 categories of safety issues and 6 user-oriented long-context tasks, with a total of 1,543 test cases,"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2502.16971","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-24T08:54:39Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b176f910cbc7823f2a36938b9c0e89a9a4b423599060f2f36338aa037fa8daa1","abstract_canon_sha256":"0ade174de7a9f01fc5287f51b3cbc15ac05408eb973a646b6aa110bcfa3f5e80"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:19:04.253919Z","signature_b64":"u5Ps853AQHzQsULb09BTcTprh2NgMoBoMT1PA8JuAcA5oixaA2bd4OEpi03Tm9kL+3InF4cO3Z7bs3nZwCXVCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fc0ce66cd95445b5af91783d4693db02b962913d5ca7a58fac4a2dd6a872f9c6","last_reissued_at":"2026-07-05T10:19:04.253404Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:19:04.253404Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LongSafety: Evaluating Long-Context Safety of Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Cunxiang Wang, Hongning Wang, Jiale Cheng, Jie Tang, Minlie Huang, Shiyao Cui, Xiaotao Gu, Yida Lu, Yuxiao Dong, Zhexin Zhang","submitted_at":"2025-02-24T08:54:39Z","abstract_excerpt":"As Large Language Models (LLMs) continue to advance in understanding and generating long sequences, new safety concerns have been introduced through the long context. However, the safety of LLMs in long-context tasks remains under-explored, leaving a significant gap in both evaluation and improvement of their safety. To address this, we introduce LongSafety, the first comprehensive benchmark specifically designed to evaluate LLM safety in open-ended long-context tasks. LongSafety encompasses 7 categories of safety issues and 6 user-oriented long-context tasks, with a total of 1,543 test cases,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.16971","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/2502.16971/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2502.16971","created_at":"2026-07-05T10:19:04.253471+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.16971v1","created_at":"2026-07-05T10:19:04.253471+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.16971","created_at":"2026-07-05T10:19:04.253471+00:00"},{"alias_kind":"pith_short_12","alias_value":"7QGOM3GZKRC3","created_at":"2026-07-05T10:19:04.253471+00:00"},{"alias_kind":"pith_short_16","alias_value":"7QGOM3GZKRC3LL4R","created_at":"2026-07-05T10:19:04.253471+00:00"},{"alias_kind":"pith_short_8","alias_value":"7QGOM3GZ","created_at":"2026-07-05T10:19:04.253471+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19887","citing_title":"FinRED: An Expert-Guided Benchmark Generation and Evaluation Framework for Financial LLM Red-Teaming","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2512.21110","citing_title":"Beyond Context: Large Language Models' Failure to Grasp Users' Intent","ref_index":75,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK","json":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK.json","graph_json":"https://pith.science/api/pith-number/7QGOM3GZKRC3LL4RPA6UNE63AK/graph.json","events_json":"https://pith.science/api/pith-number/7QGOM3GZKRC3LL4RPA6UNE63AK/events.json","paper":"https://pith.science/paper/7QGOM3GZ"},"agent_actions":{"view_html":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK","download_json":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK.json","view_paper":"https://pith.science/paper/7QGOM3GZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.16971&json=true","fetch_graph":"https://pith.science/api/pith-number/7QGOM3GZKRC3LL4RPA6UNE63AK/graph.json","fetch_events":"https://pith.science/api/pith-number/7QGOM3GZKRC3LL4RPA6UNE63AK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK/action/storage_attestation","attest_author":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK/action/author_attestation","sign_citation":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK/action/citation_signature","submit_replication":"https://pith.science/pith/7QGOM3GZKRC3LL4RPA6UNE63AK/action/replication_record"}},"created_at":"2026-07-05T10:19:04.253471+00:00","updated_at":"2026-07-05T10:19:04.253471+00:00"}