{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:444SPTCJWX2NVXKZ4FNYJGLLJL","short_pith_number":"pith:444SPTCJ","schema_version":"1.0","canonical_sha256":"e73927cc49b5f4dadd59e15b84996b4afb61ae524f9b44a91f8726d742522fa6","source":{"kind":"arxiv","id":"2203.09509","version":4},"attestation_state":"computed","paper":{"title":"ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dipankar Ray, Ece Kamar, Hamid Palangi, Maarten Sap, Saadia Gabriel, Thomas Hartvigsen","submitted_at":"2022-03-17T17:57:56Z","abstract_excerpt":"Toxic language detection systems often falsely flag text that contains minority group mentions as toxic, as those groups are often the targets of online hate. Such over-reliance on spurious correlations also causes systems to struggle with detecting implicitly toxic language. To help mitigate these issues, we create ToxiGen, a new large-scale and machine-generated dataset of 274k toxic and benign statements about 13 minority groups. We develop a demonstration-based prompting framework and an adversarial classifier-in-the-loop decoding method to generate subtly toxic and benign text with a mass"},"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":"2203.09509","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-17T17:57:56Z","cross_cats_sorted":[],"title_canon_sha256":"7cf7da14ecd3b7c4e7a5cbc0ef211edf7a35f2ea54e1b6d2d355568258e7f29f","abstract_canon_sha256":"1c4787f567c048c0da6ea3952d7f6cbb8e2fd41997cdc03e8b599e9b1d6149df"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:40:10.040145Z","signature_b64":"IY5143SpQKRAZV98VU4vZyKI0pmwf3J+IINA12vVInYGO+phfmfAEvqD0YajK+US3EZjd8nE6pMkGunpYd8GDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e73927cc49b5f4dadd59e15b84996b4afb61ae524f9b44a91f8726d742522fa6","last_reissued_at":"2026-07-05T04:40:10.039665Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:40:10.039665Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dipankar Ray, Ece Kamar, Hamid Palangi, Maarten Sap, Saadia Gabriel, Thomas Hartvigsen","submitted_at":"2022-03-17T17:57:56Z","abstract_excerpt":"Toxic language detection systems often falsely flag text that contains minority group mentions as toxic, as those groups are often the targets of online hate. Such over-reliance on spurious correlations also causes systems to struggle with detecting implicitly toxic language. To help mitigate these issues, we create ToxiGen, a new large-scale and machine-generated dataset of 274k toxic and benign statements about 13 minority groups. We develop a demonstration-based prompting framework and an adversarial classifier-in-the-loop decoding method to generate subtly toxic and benign text with a mass"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.09509","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/2203.09509/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":"2203.09509","created_at":"2026-07-05T04:40:10.039723+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.09509v4","created_at":"2026-07-05T04:40:10.039723+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.09509","created_at":"2026-07-05T04:40:10.039723+00:00"},{"alias_kind":"pith_short_12","alias_value":"444SPTCJWX2N","created_at":"2026-07-05T04:40:10.039723+00:00"},{"alias_kind":"pith_short_16","alias_value":"444SPTCJWX2NVXKZ","created_at":"2026-07-05T04:40:10.039723+00:00"},{"alias_kind":"pith_short_8","alias_value":"444SPTCJ","created_at":"2026-07-05T04:40:10.039723+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":25,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.09388","citing_title":"Distilling Safe LLM Systems via Soft Prompts for On Device Settings","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06342","citing_title":"Symmetric Divergence and Normalized Similarity: A Unified Topological Framework for Representation Analysis","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23974","citing_title":"AERIC: Anticipatory Hidden-State Monitoring for Implicit Harmful Dialogue","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25492","citing_title":"SafetyRepro: Configuration-Conditional Rank Instability on Alignment Benchmarks","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30076","citing_title":"UniSteer: Text-Guided Flow Matching in Activation Space for Versatile LLM Steering","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22643","citing_title":"Boiling the Frog: A Multi-Turn Benchmark for Agentic Safety","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2309.10305","citing_title":"Baichuan 2: Open Large-scale Language Models","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2502.01941","citing_title":"Semantic Integrity Matters: Benchmarking and Preserving High-Density Reasoning in KV Cache Compression","ref_index":73,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22643","citing_title":"Boiling the Frog: A Multi-Turn Benchmark for Agentic Safety","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2510.07239","citing_title":"Red-Bandit: Test-Time Adaptation for LLM Red-Teaming via Bandit-Guided LoRA Experts","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2401.05561","citing_title":"TrustLLM: Trustworthiness in Large Language Models","ref_index":249,"is_internal_anchor":false},{"citing_arxiv_id":"2602.11157","citing_title":"Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2512.08104","citing_title":"AgentCrypt: Advancing Privacy and (Secure) Computation in AI Agent Collaboration","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2304.06364","citing_title":"AGIEval: A Human-Centric Benchmark for Evaluating Foundation Models","ref_index":56,"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":8,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03242","citing_title":"DRAFT: Task Decoupled Latent Reasoning for Agent Safety","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2602.24176","citing_title":"Beyond Explainable AI (XAI): An Overdue Paradigm Shift and Post-XAI Research Directions","ref_index":199,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14350","citing_title":"Distributionally Robust Multi-Task Reinforcement Learning via Adaptive Task Sampling","ref_index":252,"is_internal_anchor":false},{"citing_arxiv_id":"2309.05463","citing_title":"Textbooks Are All You Need II: phi-1.5 technical report","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2310.03684","citing_title":"SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11217","citing_title":"Leveraging RAG for Training-Free Alignment of LLMs","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08116","citing_title":"The Safety-Aware Denoiser for Text Diffusion Models","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10639","citing_title":"Navigating the Sea of LLM Evaluation: Investigating Bias in Toxicity Benchmarks","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07879","citing_title":"FlowGuard: Towards Lightweight In-Generation Safety Detection for Diffusion Models via Linear Latent Decoding","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17730","citing_title":"MHSafeEval: Role-Aware Interaction-Level Evaluation of Mental Health Safety in Large Language Models","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL","json":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL.json","graph_json":"https://pith.science/api/pith-number/444SPTCJWX2NVXKZ4FNYJGLLJL/graph.json","events_json":"https://pith.science/api/pith-number/444SPTCJWX2NVXKZ4FNYJGLLJL/events.json","paper":"https://pith.science/paper/444SPTCJ"},"agent_actions":{"view_html":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL","download_json":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL.json","view_paper":"https://pith.science/paper/444SPTCJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.09509&json=true","fetch_graph":"https://pith.science/api/pith-number/444SPTCJWX2NVXKZ4FNYJGLLJL/graph.json","fetch_events":"https://pith.science/api/pith-number/444SPTCJWX2NVXKZ4FNYJGLLJL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL/action/storage_attestation","attest_author":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL/action/author_attestation","sign_citation":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL/action/citation_signature","submit_replication":"https://pith.science/pith/444SPTCJWX2NVXKZ4FNYJGLLJL/action/replication_record"}},"created_at":"2026-07-05T04:40:10.039723+00:00","updated_at":"2026-07-05T04:40:10.039723+00:00"}