{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:E65HEJQ2MA6MJ3MICIFJLOSU3C","short_pith_number":"pith:E65HEJQ2","schema_version":"1.0","canonical_sha256":"27ba72261a603cc4ed88120a95ba54d8bb9b16d0ff81a2513b422268d6878581","source":{"kind":"arxiv","id":"2402.05044","version":4},"attestation_state":"computed","paper":{"title":"SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bowen Dong, Dahua Lin, Jing Shao, Lijun Li, Ruohui Wang, Wangmeng Zuo, Xuhao Hu, Yu Qiao","submitted_at":"2024-02-07T17:33:54Z","abstract_excerpt":"In the rapidly evolving landscape of Large Language Models (LLMs), ensuring robust safety measures is paramount. To meet this crucial need, we propose \\emph{SALAD-Bench}, a safety benchmark specifically designed for evaluating LLMs, attack, and defense methods. Distinguished by its breadth, SALAD-Bench transcends conventional benchmarks through its large scale, rich diversity, intricate taxonomy spanning three levels, and versatile functionalities.SALAD-Bench is crafted with a meticulous array of questions, from standard queries to complex ones enriched with attack, defense modifications and m"},"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":"2402.05044","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-02-07T17:33:54Z","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"title_canon_sha256":"bbd0bc667c2fa282598628a6dc43b79b63a40a302d687d02ad766d733926260d","abstract_canon_sha256":"5b948a8b93d5c8a512c14a0de3ebf830cc2896ce9dd68e98bd9a8a2638ea67d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:37.900089Z","signature_b64":"Hz2EzAO3q2QI9isim8MyUOwncWSf9FnpBEQ+iwRM/8IROxIPByFNH6bj2pgir7cVfAkM5vkEHMYy4oXktb/EBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"27ba72261a603cc4ed88120a95ba54d8bb9b16d0ff81a2513b422268d6878581","last_reissued_at":"2026-07-05T08:28:37.899666Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:37.899666Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bowen Dong, Dahua Lin, Jing Shao, Lijun Li, Ruohui Wang, Wangmeng Zuo, Xuhao Hu, Yu Qiao","submitted_at":"2024-02-07T17:33:54Z","abstract_excerpt":"In the rapidly evolving landscape of Large Language Models (LLMs), ensuring robust safety measures is paramount. To meet this crucial need, we propose \\emph{SALAD-Bench}, a safety benchmark specifically designed for evaluating LLMs, attack, and defense methods. Distinguished by its breadth, SALAD-Bench transcends conventional benchmarks through its large scale, rich diversity, intricate taxonomy spanning three levels, and versatile functionalities.SALAD-Bench is crafted with a meticulous array of questions, from standard queries to complex ones enriched with attack, defense modifications and m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.05044","kind":"arxiv","version":4},"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/2402.05044/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":"2402.05044","created_at":"2026-07-05T08:28:37.899727+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.05044v4","created_at":"2026-07-05T08:28:37.899727+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.05044","created_at":"2026-07-05T08:28:37.899727+00:00"},{"alias_kind":"pith_short_12","alias_value":"E65HEJQ2MA6M","created_at":"2026-07-05T08:28:37.899727+00:00"},{"alias_kind":"pith_short_16","alias_value":"E65HEJQ2MA6MJ3MI","created_at":"2026-07-05T08:28:37.899727+00:00"},{"alias_kind":"pith_short_8","alias_value":"E65HEJQ2","created_at":"2026-07-05T08:28:37.899727+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":22,"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":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02530","citing_title":"SafeSteer: Localized On-Policy Distillation for Efficient Safety Alignment","ref_index":74,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02959","citing_title":"Gate AI: LLM Security Benchmark Evaluation Methodology and Results","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00160","citing_title":"DataShield: Safety-degrading Data Filtering for LLM Benign Instruction Fine-Tuning","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28153","citing_title":"Robust Harmful Features Under Jailbreak Attacks: Mechanistic Evidence from Attention Head Specialization in Large Language Models","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03226","citing_title":"Self-Mined Hardness for Safety Fine-Tuning","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14749","citing_title":"Non-linear Interventions on Large Language Models","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.29068","citing_title":"Robust and Efficient Guardrails with Latent Reasoning","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.29659","citing_title":"Opir: Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2602.07340","citing_title":"Revisiting Robustness for LLM Safety Alignment via Selective Geometry Control","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2511.12710","citing_title":"Evolve the Method, Not the Prompts: Evolutionary Synthesis of Jailbreak Attacks on LLMs","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2407.21772","citing_title":"ShieldGemma: Generative AI Content Moderation Based on Gemma","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17329","citing_title":"LPG: Balancing Efficiency and Policy Reasoning in Latent Policy Guardrails","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2406.18495","citing_title":"WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2602.05946","citing_title":"f-GRPO and Beyond: Divergence-Based Reinforcement Learning Algorithms for General LLM Alignment","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10639","citing_title":"Navigating the Sea of LLM Evaluation: Investigating Bias in Toxicity Benchmarks","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24074","citing_title":"How Sensitive Are Safety Benchmarks to Judge Configuration Choices?","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00245","citing_title":"ARMOR 2025: A Military-Aligned Benchmark for Evaluating Large Language Model Safety Beyond Civilian Contexts","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07754","citing_title":"The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18756","citing_title":"Towards Understanding the Robustness of Sparse Autoencoders","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03179","citing_title":"A Validated Prompt Bank for Malicious Code Generation: Separating Executable Weapons from Security Knowledge in 1,554 Consensus-Labeled Prompts","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03226","citing_title":"Self-Mined Hardness for Safety Fine-Tuning","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C","json":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C.json","graph_json":"https://pith.science/api/pith-number/E65HEJQ2MA6MJ3MICIFJLOSU3C/graph.json","events_json":"https://pith.science/api/pith-number/E65HEJQ2MA6MJ3MICIFJLOSU3C/events.json","paper":"https://pith.science/paper/E65HEJQ2"},"agent_actions":{"view_html":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C","download_json":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C.json","view_paper":"https://pith.science/paper/E65HEJQ2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.05044&json=true","fetch_graph":"https://pith.science/api/pith-number/E65HEJQ2MA6MJ3MICIFJLOSU3C/graph.json","fetch_events":"https://pith.science/api/pith-number/E65HEJQ2MA6MJ3MICIFJLOSU3C/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C/action/timestamp_anchor","attest_storage":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C/action/storage_attestation","attest_author":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C/action/author_attestation","sign_citation":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C/action/citation_signature","submit_replication":"https://pith.science/pith/E65HEJQ2MA6MJ3MICIFJLOSU3C/action/replication_record"}},"created_at":"2026-07-05T08:28:37.899727+00:00","updated_at":"2026-07-05T08:28:37.899727+00:00"}