{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:WWVJAGVIL4CHPE247U25J7GJIB","short_pith_number":"pith:WWVJAGVI","canonical_record":{"source":{"id":"2407.13928","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-18T22:32:20Z","cross_cats_sorted":[],"title_canon_sha256":"ad9e74c92e5e49315f8971292f7193803f21d901f84d030d11ce5e5f30c32ec5","abstract_canon_sha256":"430e2f76597f512134091f9025cf89f62ad2d42ba9c4a59795f42482983b6cec"},"schema_version":"1.0"},"canonical_sha256":"b5aa901aa85f0477935cfd35d4fcc940543106a43a13f3d3fbbc3cf2f7969161","source":{"kind":"arxiv","id":"2407.13928","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.13928","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"arxiv_version","alias_value":"2407.13928v1","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.13928","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"pith_short_12","alias_value":"WWVJAGVIL4CH","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"pith_short_16","alias_value":"WWVJAGVIL4CHPE24","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"pith_short_8","alias_value":"WWVJAGVI","created_at":"2026-07-05T08:45:55Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:WWVJAGVIL4CHPE247U25J7GJIB","target":"record","payload":{"canonical_record":{"source":{"id":"2407.13928","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-18T22:32:20Z","cross_cats_sorted":[],"title_canon_sha256":"ad9e74c92e5e49315f8971292f7193803f21d901f84d030d11ce5e5f30c32ec5","abstract_canon_sha256":"430e2f76597f512134091f9025cf89f62ad2d42ba9c4a59795f42482983b6cec"},"schema_version":"1.0"},"canonical_sha256":"b5aa901aa85f0477935cfd35d4fcc940543106a43a13f3d3fbbc3cf2f7969161","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:45:55.655097Z","signature_b64":"22skYYtVX1csicC1driuRrTqYFmrvxccaM9+CdQpLvdKkKNZjwy9USV9/8sfcRk8Y91nfJWF/VRMf0Qh7PPgBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b5aa901aa85f0477935cfd35d4fcc940543106a43a13f3d3fbbc3cf2f7969161","last_reissued_at":"2026-07-05T08:45:55.654559Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:45:55.654559Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2407.13928","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:45:55Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NyIoXi0S8IIaPh4Zm2tiTZpsglE/f90lE+DBZCNrsf43ReblHv8Lvp8E32TprXoz53dsS2wktOCysli6eQyZBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T06:22:59.445194Z"},"content_sha256":"f598d08c7919f41903b3957a482114b3f1556bc64e15837b4f1780bb54d1dc2e","schema_version":"1.0","event_id":"sha256:f598d08c7919f41903b3957a482114b3f1556bc64e15837b4f1780bb54d1dc2e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:WWVJAGVIL4CHPE247U25J7GJIB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"BiasDPO: Mitigating Bias in Language Models through Direct Preference Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ahmed Allam","submitted_at":"2024-07-18T22:32:20Z","abstract_excerpt":"Large Language Models (LLMs) have become pivotal in advancing natural language processing, yet their potential to perpetuate biases poses significant concerns. This paper introduces a new framework employing Direct Preference Optimization (DPO) to mitigate gender, racial, and religious biases in LLM-generated English text. By developing a loss function that favors less biased over biased completions, our approach cultivates a preference for respectful and non-discriminatory language in LLMs. We also contribute a manually designed dataset for training LLMs to recognize and correct biases. This "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.13928","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/2407.13928/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:45:55Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/2/TIbxftdSP668ilGXO7rryRbOfa5489cFQYS3Fkvc+Wnauad51duONUB9kw/Cl1nL4aJ76FFsU8NKRA9LxDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T06:22:59.445754Z"},"content_sha256":"fdd1e66f169906239d8a99e677eb53e11f806c642901a25d2404d79d5341f33b","schema_version":"1.0","event_id":"sha256:fdd1e66f169906239d8a99e677eb53e11f806c642901a25d2404d79d5341f33b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/WWVJAGVIL4CHPE247U25J7GJIB/bundle.json","state_url":"https://pith.science/pith/WWVJAGVIL4CHPE247U25J7GJIB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/WWVJAGVIL4CHPE247U25J7GJIB/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-07T06:22:59Z","links":{"resolver":"https://pith.science/pith/WWVJAGVIL4CHPE247U25J7GJIB","bundle":"https://pith.science/pith/WWVJAGVIL4CHPE247U25J7GJIB/bundle.json","state":"https://pith.science/pith/WWVJAGVIL4CHPE247U25J7GJIB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/WWVJAGVIL4CHPE247U25J7GJIB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:WWVJAGVIL4CHPE247U25J7GJIB","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"430e2f76597f512134091f9025cf89f62ad2d42ba9c4a59795f42482983b6cec","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-18T22:32:20Z","title_canon_sha256":"ad9e74c92e5e49315f8971292f7193803f21d901f84d030d11ce5e5f30c32ec5"},"schema_version":"1.0","source":{"id":"2407.13928","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.13928","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"arxiv_version","alias_value":"2407.13928v1","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.13928","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"pith_short_12","alias_value":"WWVJAGVIL4CH","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"pith_short_16","alias_value":"WWVJAGVIL4CHPE24","created_at":"2026-07-05T08:45:55Z"},{"alias_kind":"pith_short_8","alias_value":"WWVJAGVI","created_at":"2026-07-05T08:45:55Z"}],"graph_snapshots":[{"event_id":"sha256:fdd1e66f169906239d8a99e677eb53e11f806c642901a25d2404d79d5341f33b","target":"graph","created_at":"2026-07-05T08:45:55Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2407.13928/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) have become pivotal in advancing natural language processing, yet their potential to perpetuate biases poses significant concerns. This paper introduces a new framework employing Direct Preference Optimization (DPO) to mitigate gender, racial, and religious biases in LLM-generated English text. By developing a loss function that favors less biased over biased completions, our approach cultivates a preference for respectful and non-discriminatory language in LLMs. We also contribute a manually designed dataset for training LLMs to recognize and correct biases. This ","authors_text":"Ahmed Allam","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-18T22:32:20Z","title":"BiasDPO: Mitigating Bias in Language Models through Direct Preference Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.13928","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:f598d08c7919f41903b3957a482114b3f1556bc64e15837b4f1780bb54d1dc2e","target":"record","created_at":"2026-07-05T08:45:55Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"430e2f76597f512134091f9025cf89f62ad2d42ba9c4a59795f42482983b6cec","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-18T22:32:20Z","title_canon_sha256":"ad9e74c92e5e49315f8971292f7193803f21d901f84d030d11ce5e5f30c32ec5"},"schema_version":"1.0","source":{"id":"2407.13928","kind":"arxiv","version":1}},"canonical_sha256":"b5aa901aa85f0477935cfd35d4fcc940543106a43a13f3d3fbbc3cf2f7969161","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b5aa901aa85f0477935cfd35d4fcc940543106a43a13f3d3fbbc3cf2f7969161","first_computed_at":"2026-07-05T08:45:55.654559Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:45:55.654559Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"22skYYtVX1csicC1driuRrTqYFmrvxccaM9+CdQpLvdKkKNZjwy9USV9/8sfcRk8Y91nfJWF/VRMf0Qh7PPgBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:45:55.655097Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.13928","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f598d08c7919f41903b3957a482114b3f1556bc64e15837b4f1780bb54d1dc2e","sha256:fdd1e66f169906239d8a99e677eb53e11f806c642901a25d2404d79d5341f33b"],"state_sha256":"ad58424dee1945a949e4ecfab5d6bf376b8ffed871c7343eabf660315fafa188"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"d11zphi9g92q+/k94nZjMUp8/WHKMZfJHRPQYqeWh+l3MPPdZUDS+4JDJgc5+fX4P7eyFiTcGEH4OfPWr0oJDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T06:22:59.449234Z","bundle_sha256":"ca093ff44b1d03a9d287f39b8bdec6d8574f6a2869bc40bcc302fed990d1ac8c"}}