{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:F4XQ3GQQ5KYQ5WAD5M4IECPJZV","short_pith_number":"pith:F4XQ3GQQ","schema_version":"1.0","canonical_sha256":"2f2f0d9a10eab10ed803eb388209e9cd654c152d9601e0460974cc84d7c295f8","source":{"kind":"arxiv","id":"2202.11176","version":1},"attestation_state":"computed","paper":{"title":"A New Generation of Perspective API: Efficient Multilingual Character-level Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CY","cs.LG"],"primary_cat":"cs.CL","authors_text":"Alyssa Lees, Donald Metzler, Jai Gupta, Jeffrey Sorensen, Lucy Vasserman, Vinh Q. Tran, Yi Tay","submitted_at":"2022-02-22T20:55:31Z","abstract_excerpt":"On the world wide web, toxic content detectors are a crucial line of defense against potentially hateful and offensive messages. As such, building highly effective classifiers that enable a safer internet is an important research area. Moreover, the web is a highly multilingual, cross-cultural community that develops its own lingo over time. As such, it is crucial to develop models that are effective across a diverse range of languages, usages, and styles. In this paper, we present the fundamentals behind the next version of the Perspective API from Google Jigsaw. At the heart of the approach "},"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":"2202.11176","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-02-22T20:55:31Z","cross_cats_sorted":["cs.AI","cs.CY","cs.LG"],"title_canon_sha256":"e9503e69bb12d121b4320b823940dd81763b28806e75f2c57112290a769f4666","abstract_canon_sha256":"4f230da25b10585f4a093c0326ca57e7b0a532b07abef011b9178a74e44bf886"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:59:29.152545Z","signature_b64":"drqOvWTeeU2aTvQih+fe9wzIrHLn47eiydks/EIw+9wthiWZDh6czvWAlfUlp2zeLnsjFttJ7R7sS/LSBiSoDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2f2f0d9a10eab10ed803eb388209e9cd654c152d9601e0460974cc84d7c295f8","last_reissued_at":"2026-07-05T03:59:29.152195Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:59:29.152195Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A New Generation of Perspective API: Efficient Multilingual Character-level Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CY","cs.LG"],"primary_cat":"cs.CL","authors_text":"Alyssa Lees, Donald Metzler, Jai Gupta, Jeffrey Sorensen, Lucy Vasserman, Vinh Q. Tran, Yi Tay","submitted_at":"2022-02-22T20:55:31Z","abstract_excerpt":"On the world wide web, toxic content detectors are a crucial line of defense against potentially hateful and offensive messages. As such, building highly effective classifiers that enable a safer internet is an important research area. Moreover, the web is a highly multilingual, cross-cultural community that develops its own lingo over time. As such, it is crucial to develop models that are effective across a diverse range of languages, usages, and styles. In this paper, we present the fundamentals behind the next version of the Perspective API from Google Jigsaw. At the heart of the approach "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.11176","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/2202.11176/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":"2202.11176","created_at":"2026-07-05T03:59:29.152250+00:00"},{"alias_kind":"arxiv_version","alias_value":"2202.11176v1","created_at":"2026-07-05T03:59:29.152250+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.11176","created_at":"2026-07-05T03:59:29.152250+00:00"},{"alias_kind":"pith_short_12","alias_value":"F4XQ3GQQ5KYQ","created_at":"2026-07-05T03:59:29.152250+00:00"},{"alias_kind":"pith_short_16","alias_value":"F4XQ3GQQ5KYQ5WAD","created_at":"2026-07-05T03:59:29.152250+00:00"},{"alias_kind":"pith_short_8","alias_value":"F4XQ3GQQ","created_at":"2026-07-05T03:59:29.152250+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26947","citing_title":"KZ-SafetyPrompts: A Kazakh Safety Evaluation Prompt Dataset for Large Language Models","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17187","citing_title":"PluRule: A Benchmark for Moderating Pluralistic Communities on Social Media","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2509.18127","citing_title":"Safe-SAIL: Towards a Fine-grained Safety Landscape of Large Language Models via Sparse Autoencoder Interpretation Framework","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2412.03555","citing_title":"PaliGemma 2: A Family of Versatile VLMs for Transfer","ref_index":46,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV","json":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV.json","graph_json":"https://pith.science/api/pith-number/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/graph.json","events_json":"https://pith.science/api/pith-number/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/events.json","paper":"https://pith.science/paper/F4XQ3GQQ"},"agent_actions":{"view_html":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV","download_json":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV.json","view_paper":"https://pith.science/paper/F4XQ3GQQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2202.11176&json=true","fetch_graph":"https://pith.science/api/pith-number/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/graph.json","fetch_events":"https://pith.science/api/pith-number/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/action/storage_attestation","attest_author":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/action/author_attestation","sign_citation":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/action/citation_signature","submit_replication":"https://pith.science/pith/F4XQ3GQQ5KYQ5WAD5M4IECPJZV/action/replication_record"}},"created_at":"2026-07-05T03:59:29.152250+00:00","updated_at":"2026-07-05T03:59:29.152250+00:00"}