{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:UJAAJ76QORXPNPAR73HQP6DDXO","short_pith_number":"pith:UJAAJ76Q","schema_version":"1.0","canonical_sha256":"a24004ffd0746ef6bc11fecf07f863bbac7d909f2979c0288162ac64e92808fe","source":{"kind":"arxiv","id":"2005.02439","version":3},"attestation_state":"computed","paper":{"title":"Contextualizing Hate Speech Classifiers with Post-hoc Explanation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aida Mostafazadeh Davani, Brendan Kennedy, Morteza Dehghani, Xiang Ren, Xisen Jin","submitted_at":"2020-05-05T18:56:40Z","abstract_excerpt":"Hate speech classifiers trained on imbalanced datasets struggle to determine if group identifiers like \"gay\" or \"black\" are used in offensive or prejudiced ways. Such biases manifest in false positives when these identifiers are present, due to models' inability to learn the contexts which constitute a hateful usage of identifiers. We extract SOC post-hoc explanations from fine-tuned BERT classifiers to efficiently detect bias towards identity terms. Then, we propose a novel regularization technique based on these explanations that encourages models to learn from the context of group identifie"},"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":"2005.02439","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-05-05T18:56:40Z","cross_cats_sorted":["cs.IR","cs.LG"],"title_canon_sha256":"f35336434896e6bb94b736017b9dde2caa150990db686856756c01a4386e17a3","abstract_canon_sha256":"1583dab3e18d8559f6c9cd47d6a621918f835a021bdd587ff4664a6692fafc88"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:16:50.343693Z","signature_b64":"k6NmIV8TR5y4gM3d7FKE0gjBK8ljreohuGuGhBkIF1PwNTAAUVR3m1Dp2sS2KDNFnI6PZyIDwMvDCt2/uo7iBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a24004ffd0746ef6bc11fecf07f863bbac7d909f2979c0288162ac64e92808fe","last_reissued_at":"2026-07-05T01:16:50.343246Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:16:50.343246Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Contextualizing Hate Speech Classifiers with Post-hoc Explanation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aida Mostafazadeh Davani, Brendan Kennedy, Morteza Dehghani, Xiang Ren, Xisen Jin","submitted_at":"2020-05-05T18:56:40Z","abstract_excerpt":"Hate speech classifiers trained on imbalanced datasets struggle to determine if group identifiers like \"gay\" or \"black\" are used in offensive or prejudiced ways. Such biases manifest in false positives when these identifiers are present, due to models' inability to learn the contexts which constitute a hateful usage of identifiers. We extract SOC post-hoc explanations from fine-tuned BERT classifiers to efficiently detect bias towards identity terms. Then, we propose a novel regularization technique based on these explanations that encourages models to learn from the context of group identifie"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2005.02439","kind":"arxiv","version":3},"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/2005.02439/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":"2005.02439","created_at":"2026-07-05T01:16:50.343301+00:00"},{"alias_kind":"arxiv_version","alias_value":"2005.02439v3","created_at":"2026-07-05T01:16:50.343301+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2005.02439","created_at":"2026-07-05T01:16:50.343301+00:00"},{"alias_kind":"pith_short_12","alias_value":"UJAAJ76QORXP","created_at":"2026-07-05T01:16:50.343301+00:00"},{"alias_kind":"pith_short_16","alias_value":"UJAAJ76QORXPNPAR","created_at":"2026-07-05T01:16:50.343301+00:00"},{"alias_kind":"pith_short_8","alias_value":"UJAAJ76Q","created_at":"2026-07-05T01:16:50.343301+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.00802","citing_title":"Explanations as Bias Detectors: A Critical Study of Local Post-hoc XAI Methods for Fairness Exploration","ref_index":31,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO","json":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO.json","graph_json":"https://pith.science/api/pith-number/UJAAJ76QORXPNPAR73HQP6DDXO/graph.json","events_json":"https://pith.science/api/pith-number/UJAAJ76QORXPNPAR73HQP6DDXO/events.json","paper":"https://pith.science/paper/UJAAJ76Q"},"agent_actions":{"view_html":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO","download_json":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO.json","view_paper":"https://pith.science/paper/UJAAJ76Q","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2005.02439&json=true","fetch_graph":"https://pith.science/api/pith-number/UJAAJ76QORXPNPAR73HQP6DDXO/graph.json","fetch_events":"https://pith.science/api/pith-number/UJAAJ76QORXPNPAR73HQP6DDXO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO/action/storage_attestation","attest_author":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO/action/author_attestation","sign_citation":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO/action/citation_signature","submit_replication":"https://pith.science/pith/UJAAJ76QORXPNPAR73HQP6DDXO/action/replication_record"}},"created_at":"2026-07-05T01:16:50.343301+00:00","updated_at":"2026-07-05T01:16:50.343301+00:00"}