{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:VZBCVPYRBPUZ5IDEWVAWDGFZ3J","short_pith_number":"pith:VZBCVPYR","schema_version":"1.0","canonical_sha256":"ae422abf110be99ea064b5416198b9da685a6e1c035aa4142e485a447b855412","source":{"kind":"arxiv","id":"2208.03274","version":2},"attestation_state":"computed","paper":{"title":"A Holistic Approach to Undesired Content Detection in the Real World","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Angela Jiang, Chong Zhang, Lilian Weng, Sandhini Agarwal, Steven Adler, Teddy Lee, Todor Markov, Tyna Eloundou","submitted_at":"2022-08-05T16:47:23Z","abstract_excerpt":"We present a holistic approach to building a robust and useful natural language classification system for real-world content moderation. The success of such a system relies on a chain of carefully designed and executed steps, including the design of content taxonomies and labeling instructions, data quality control, an active learning pipeline to capture rare events, and a variety of methods to make the model robust and to avoid overfitting. Our moderation system is trained to detect a broad set of categories of undesired content, including sexual content, hateful content, violence, self-harm,"},"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":"2208.03274","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-08-05T16:47:23Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"cee6e370b6156f4a3e1cffb447234a3841362d28b1968c588444ae7060938b40","abstract_canon_sha256":"eb845aa5a95b5fee725611870bedad5bcc5e21a63238d0d3866094b23440cb25"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:41:50.895690Z","signature_b64":"7+nvfpF07xPdSoih6snQ0uvha+EwMTrMUipNLrm/F4wCLCEHDkMbTlDH2TM+L+k/ea66m7Xy3bDAz9pBLwAGCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ae422abf110be99ea064b5416198b9da685a6e1c035aa4142e485a447b855412","last_reissued_at":"2026-07-05T05:41:50.895249Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:41:50.895249Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Holistic Approach to Undesired Content Detection in the Real World","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Angela Jiang, Chong Zhang, Lilian Weng, Sandhini Agarwal, Steven Adler, Teddy Lee, Todor Markov, Tyna Eloundou","submitted_at":"2022-08-05T16:47:23Z","abstract_excerpt":"We present a holistic approach to building a robust and useful natural language classification system for real-world content moderation. The success of such a system relies on a chain of carefully designed and executed steps, including the design of content taxonomies and labeling instructions, data quality control, an active learning pipeline to capture rare events, and a variety of methods to make the model robust and to avoid overfitting. Our moderation system is trained to detect a broad set of categories of undesired content, including sexual content, hateful content, violence, self-harm,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2208.03274","kind":"arxiv","version":2},"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/2208.03274/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":"2208.03274","created_at":"2026-07-05T05:41:50.895304+00:00"},{"alias_kind":"arxiv_version","alias_value":"2208.03274v2","created_at":"2026-07-05T05:41:50.895304+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2208.03274","created_at":"2026-07-05T05:41:50.895304+00:00"},{"alias_kind":"pith_short_12","alias_value":"VZBCVPYRBPUZ","created_at":"2026-07-05T05:41:50.895304+00:00"},{"alias_kind":"pith_short_16","alias_value":"VZBCVPYRBPUZ5IDE","created_at":"2026-07-05T05:41:50.895304+00:00"},{"alias_kind":"pith_short_8","alias_value":"VZBCVPYR","created_at":"2026-07-05T05:41:50.895304+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":11,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.00527","citing_title":"AI Native Games: A Survey and Roadmap","ref_index":91,"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":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25893","citing_title":"$D^2$-Monitor: Dynamic Safety Monitoring for Diffusion LLMs via Hesitation-Aware Routing","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25454","citing_title":"AI Content Moderation in Therapy Conversations","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27632","citing_title":"Yuvion LLM: An Adversarially-Aware Large Language Model for Content And AI Safety","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2506.01770","citing_title":"ReGA: Model-Based Safeguard for LLMs via Representation-Guided Abstraction","ref_index":87,"is_internal_anchor":false},{"citing_arxiv_id":"2408.00724","citing_title":"Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models","ref_index":128,"is_internal_anchor":false},{"citing_arxiv_id":"2602.11157","citing_title":"Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.26167","citing_title":"Test-Time Safety Alignment","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2211.09527","citing_title":"Ignore Previous Prompt: Attack Techniques For Language Models","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14548","citing_title":"VoxSafeBench: Not Just What Is Said, but Who, How, and Where","ref_index":19,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J","json":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J.json","graph_json":"https://pith.science/api/pith-number/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/graph.json","events_json":"https://pith.science/api/pith-number/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/events.json","paper":"https://pith.science/paper/VZBCVPYR"},"agent_actions":{"view_html":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J","download_json":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J.json","view_paper":"https://pith.science/paper/VZBCVPYR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2208.03274&json=true","fetch_graph":"https://pith.science/api/pith-number/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/graph.json","fetch_events":"https://pith.science/api/pith-number/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/action/storage_attestation","attest_author":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/action/author_attestation","sign_citation":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/action/citation_signature","submit_replication":"https://pith.science/pith/VZBCVPYRBPUZ5IDEWVAWDGFZ3J/action/replication_record"}},"created_at":"2026-07-05T05:41:50.895304+00:00","updated_at":"2026-07-05T05:41:50.895304+00:00"}