{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:W25QNIZDKAUN6D66QQCLSSQAGM","short_pith_number":"pith:W25QNIZD","schema_version":"1.0","canonical_sha256":"b6bb06a3235028df0fde8404b94a00330859eea632509b9216d8846404fdbdad","source":{"kind":"arxiv","id":"2304.05335","version":1},"attestation_state":"computed","paper":{"title":"Toxicity in ChatGPT: Analyzing Persona-assigned Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Ameet Deshpande, Ashwin Kalyan, Karthik Narasimhan, Tanmay Rajpurohit, Vishvak Murahari","submitted_at":"2023-04-11T16:53:54Z","abstract_excerpt":"Large language models (LLMs) have shown incredible capabilities and transcended the natural language processing (NLP) community, with adoption throughout many services like healthcare, therapy, education, and customer service. Since users include people with critical information needs like students or patients engaging with chatbots, the safety of these systems is of prime importance. Therefore, a clear understanding of the capabilities and limitations of LLMs is necessary. To this end, we systematically evaluate toxicity in over half a million generations of ChatGPT, a popular dialogue-based "},"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":"2304.05335","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-04-11T16:53:54Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"3b9c8e71554b065f9c0ed4b19aa50d50d8fa82af9480855557e84157652c419b","abstract_canon_sha256":"2451b8bb7d76f50bed6b32fa32da0b1012a8e5f838d709c2f35b33eef36d8ac8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:59:58.487669Z","signature_b64":"bDmy8Bku1gbnLeUuLC110TdPnOJCOB1aEX91H/tYQPXrbugqvQSYIenmpQCEmIFJC5ijM3salNtu3IrPKDt9BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b6bb06a3235028df0fde8404b94a00330859eea632509b9216d8846404fdbdad","last_reissued_at":"2026-07-05T05:59:58.487242Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:59:58.487242Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Toxicity in ChatGPT: Analyzing Persona-assigned Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Ameet Deshpande, Ashwin Kalyan, Karthik Narasimhan, Tanmay Rajpurohit, Vishvak Murahari","submitted_at":"2023-04-11T16:53:54Z","abstract_excerpt":"Large language models (LLMs) have shown incredible capabilities and transcended the natural language processing (NLP) community, with adoption throughout many services like healthcare, therapy, education, and customer service. Since users include people with critical information needs like students or patients engaging with chatbots, the safety of these systems is of prime importance. Therefore, a clear understanding of the capabilities and limitations of LLMs is necessary. To this end, we systematically evaluate toxicity in over half a million generations of ChatGPT, a popular dialogue-based "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.05335","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/2304.05335/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":"2304.05335","created_at":"2026-07-05T05:59:58.487298+00:00"},{"alias_kind":"arxiv_version","alias_value":"2304.05335v1","created_at":"2026-07-05T05:59:58.487298+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.05335","created_at":"2026-07-05T05:59:58.487298+00:00"},{"alias_kind":"pith_short_12","alias_value":"W25QNIZDKAUN","created_at":"2026-07-05T05:59:58.487298+00:00"},{"alias_kind":"pith_short_16","alias_value":"W25QNIZDKAUN6D66","created_at":"2026-07-05T05:59:58.487298+00:00"},{"alias_kind":"pith_short_8","alias_value":"W25QNIZD","created_at":"2026-07-05T05:59:58.487298+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":10,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2312.03853","citing_title":"Dr. Jekyll and Mr. Hyde: Two Faces of LLMs","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2410.04155","citing_title":"Toxic Subword Pruning for Dialogue Response Generation on Large Language Models","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2401.05561","citing_title":"TrustLLM: Trustworthiness in Large Language Models","ref_index":247,"is_internal_anchor":false},{"citing_arxiv_id":"2309.10253","citing_title":"GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12288","citing_title":"TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching","ref_index":133,"is_internal_anchor":false},{"citing_arxiv_id":"2308.11432","citing_title":"A Survey on Large Language Model based Autonomous Agents","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14514","citing_title":"Defenses at Odds: Measuring and Explaining Defense Conflicts in Large Language Models","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2310.03684","citing_title":"SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12288","citing_title":"TokenRatio: Principled Token-Level Preference Optimization via Ratio Matching","ref_index":133,"is_internal_anchor":false},{"citing_arxiv_id":"2310.08419","citing_title":"Jailbreaking Black Box Large Language Models in Twenty Queries","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM","json":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM.json","graph_json":"https://pith.science/api/pith-number/W25QNIZDKAUN6D66QQCLSSQAGM/graph.json","events_json":"https://pith.science/api/pith-number/W25QNIZDKAUN6D66QQCLSSQAGM/events.json","paper":"https://pith.science/paper/W25QNIZD"},"agent_actions":{"view_html":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM","download_json":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM.json","view_paper":"https://pith.science/paper/W25QNIZD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2304.05335&json=true","fetch_graph":"https://pith.science/api/pith-number/W25QNIZDKAUN6D66QQCLSSQAGM/graph.json","fetch_events":"https://pith.science/api/pith-number/W25QNIZDKAUN6D66QQCLSSQAGM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM/action/storage_attestation","attest_author":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM/action/author_attestation","sign_citation":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM/action/citation_signature","submit_replication":"https://pith.science/pith/W25QNIZDKAUN6D66QQCLSSQAGM/action/replication_record"}},"created_at":"2026-07-05T05:59:58.487298+00:00","updated_at":"2026-07-05T05:59:58.487298+00:00"}