{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:5P3DZCLVV5F5JYST4UFW7GN6AH","short_pith_number":"pith:5P3DZCLV","schema_version":"1.0","canonical_sha256":"ebf63c8975af4bd4e253e50b6f99be01c107521f4ffecd3b54a62fd6ccaa33fe","source":{"kind":"arxiv","id":"2310.18913","version":4},"attestation_state":"computed","paper":{"title":"Debiasing Algorithm through Model Adaptation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.CL","authors_text":"David Mare\\v{c}ek, Tomasz Limisiewicz, Tom\\'a\\v{s} Musil","submitted_at":"2023-10-29T05:50:03Z","abstract_excerpt":"Large language models are becoming the go-to solution for the ever-growing number of tasks. However, with growing capacity, models are prone to rely on spurious correlations stemming from biases and stereotypes present in the training data. This work proposes a novel method for detecting and mitigating gender bias in language models. We perform causal analysis to identify problematic model components and discover that mid-upper feed-forward layers are most prone to convey bias. Based on the analysis results, we intervene in the model by applying a linear projection to the weight matrices of th"},"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":"2310.18913","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-29T05:50:03Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"23048f0b14c3d5d0470e4a56d28ac44c2942ac1e39c51917925378ded1f7d202","abstract_canon_sha256":"3df0f840be371f2441139710ae91336f81fde9a2322ca778d77d3b57b2f74816"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:24:38.186827Z","signature_b64":"Fs7PcAQyxw7cP78uRSIj/6z5DXFTxm2fE4mhP7dvt/8bRjO5qPuMn703HjVJDWnDBbrSAdO8VNrUabQhIH3FAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ebf63c8975af4bd4e253e50b6f99be01c107521f4ffecd3b54a62fd6ccaa33fe","last_reissued_at":"2026-07-05T08:24:38.186303Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:24:38.186303Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Debiasing Algorithm through Model Adaptation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.CL","authors_text":"David Mare\\v{c}ek, Tomasz Limisiewicz, Tom\\'a\\v{s} Musil","submitted_at":"2023-10-29T05:50:03Z","abstract_excerpt":"Large language models are becoming the go-to solution for the ever-growing number of tasks. However, with growing capacity, models are prone to rely on spurious correlations stemming from biases and stereotypes present in the training data. This work proposes a novel method for detecting and mitigating gender bias in language models. We perform causal analysis to identify problematic model components and discover that mid-upper feed-forward layers are most prone to convey bias. Based on the analysis results, we intervene in the model by applying a linear projection to the weight matrices of th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.18913","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/2310.18913/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":"2310.18913","created_at":"2026-07-05T08:24:38.186376+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.18913v4","created_at":"2026-07-05T08:24:38.186376+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.18913","created_at":"2026-07-05T08:24:38.186376+00:00"},{"alias_kind":"pith_short_12","alias_value":"5P3DZCLVV5F5","created_at":"2026-07-05T08:24:38.186376+00:00"},{"alias_kind":"pith_short_16","alias_value":"5P3DZCLVV5F5JYST","created_at":"2026-07-05T08:24:38.186376+00:00"},{"alias_kind":"pith_short_8","alias_value":"5P3DZCLV","created_at":"2026-07-05T08:24:38.186376+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.06052","citing_title":"Attention, May I Have Your Decision? Localizing Generative Choices in Diffusion Models","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH","json":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH.json","graph_json":"https://pith.science/api/pith-number/5P3DZCLVV5F5JYST4UFW7GN6AH/graph.json","events_json":"https://pith.science/api/pith-number/5P3DZCLVV5F5JYST4UFW7GN6AH/events.json","paper":"https://pith.science/paper/5P3DZCLV"},"agent_actions":{"view_html":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH","download_json":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH.json","view_paper":"https://pith.science/paper/5P3DZCLV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.18913&json=true","fetch_graph":"https://pith.science/api/pith-number/5P3DZCLVV5F5JYST4UFW7GN6AH/graph.json","fetch_events":"https://pith.science/api/pith-number/5P3DZCLVV5F5JYST4UFW7GN6AH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH/action/storage_attestation","attest_author":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH/action/author_attestation","sign_citation":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH/action/citation_signature","submit_replication":"https://pith.science/pith/5P3DZCLVV5F5JYST4UFW7GN6AH/action/replication_record"}},"created_at":"2026-07-05T08:24:38.186376+00:00","updated_at":"2026-07-05T08:24:38.186376+00:00"}