{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:CJEGRICZY52T4MQV7HOLIFR53R","short_pith_number":"pith:CJEGRICZ","canonical_record":{"source":{"id":"2607.28319","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-30T14:54:33Z","cross_cats_sorted":["cs.CY","cs.LG"],"title_canon_sha256":"9e5991f17b9bc6e881745e84560e72e1117517aad133feddabb79ea2201adf77","abstract_canon_sha256":"303fc48d0b46a0673ef2148621d57d51ab9f6ec75b0f53f5003b8356e88ab9cd"},"schema_version":"1.0"},"canonical_sha256":"124868a059c7753e3215f9dcb4163ddc7874fbcabf6a9fdaaca2fb94f4ddeb7a","source":{"kind":"arxiv","id":"2607.28319","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.28319","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"arxiv_version","alias_value":"2607.28319v1","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.28319","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"pith_short_12","alias_value":"CJEGRICZY52T","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"pith_short_16","alias_value":"CJEGRICZY52T4MQV","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"pith_short_8","alias_value":"CJEGRICZ","created_at":"2026-07-31T01:37:13Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:CJEGRICZY52T4MQV7HOLIFR53R","target":"record","payload":{"canonical_record":{"source":{"id":"2607.28319","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-30T14:54:33Z","cross_cats_sorted":["cs.CY","cs.LG"],"title_canon_sha256":"9e5991f17b9bc6e881745e84560e72e1117517aad133feddabb79ea2201adf77","abstract_canon_sha256":"303fc48d0b46a0673ef2148621d57d51ab9f6ec75b0f53f5003b8356e88ab9cd"},"schema_version":"1.0"},"canonical_sha256":"124868a059c7753e3215f9dcb4163ddc7874fbcabf6a9fdaaca2fb94f4ddeb7a","receipt":{"kind":"pith_receipt","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"124868a059c7753e3215f9dcb4163ddc7874fbcabf6a9fdaaca2fb94f4ddeb7a","last_reissued_at":"2026-07-31T01:37:13.756269Z","signature_status":"unsigned_v0","first_computed_at":"2026-07-31T01:37:13.756269Z"},"source_kind":"arxiv","source_id":"2607.28319","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":"2026-07-31T01:37:13Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xPc+aYXiILIP5CYW5odxcR0mwOZ9/o8lykY40uF5RohotZAfWPHy6cWSJNEXSw16k3YuXcoF7ZwJ20AfnlD2Dw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T02:42:31.427615Z"},"content_sha256":"0a7cedb4323d78869b7f0cc9c63841936db22f0aa88473668cdbc6a3f4589bd9","schema_version":"1.0","event_id":"sha256:0a7cedb4323d78869b7f0cc9c63841936db22f0aa88473668cdbc6a3f4589bd9"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:CJEGRICZY52T4MQV7HOLIFR53R","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Fairness Pruning: Locating Demographic Bias in GLU-MLP Layers via Differential Activations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CY","cs.LG"],"primary_cat":"cs.CL","authors_text":"Alfonso Ure\\~na L\\'opez, Eugenio Mart\\'inez C\\'amara, Pere Martra","submitted_at":"2026-07-30T14:54:33Z","abstract_excerpt":"This work presents Fairness Pruning, a lightweight structural intervention method designed for the management and future mitigation of demographic bias in large language models (LLMs). As a foundational empirical validation of this method, this work focuses on causal bias localization. Using minimally contrastive prompt pairs and inference-time activation capture, the method identifies neurons that react differentially when processing demographic attributes in GLU architectures, evaluating the signal at the down_proj input. Empirical evaluation was conducted on models of up to 3 billion parame"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.28319","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/2607.28319/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":"2026-07-31T01:37:13Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"M6PHTrk1HBr5NslWwSqijHbGYI9yD/YwVNzfo6DWOF0ZJPOfC60aAEiR6fENG6RXtne7Wwd9L1SH9MZGKql7AA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T02:42:31.428119Z"},"content_sha256":"cacf39edc843a1ac93ec082edfaeec58d8849ec8db16108e709c230a1792eba7","schema_version":"1.0","event_id":"sha256:cacf39edc843a1ac93ec082edfaeec58d8849ec8db16108e709c230a1792eba7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CJEGRICZY52T4MQV7HOLIFR53R/bundle.json","state_url":"https://pith.science/pith/CJEGRICZY52T4MQV7HOLIFR53R/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CJEGRICZY52T4MQV7HOLIFR53R/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":"2026-08-04T02:42:31Z","links":{"resolver":"https://pith.science/pith/CJEGRICZY52T4MQV7HOLIFR53R","bundle":"https://pith.science/pith/CJEGRICZY52T4MQV7HOLIFR53R/bundle.json","state":"https://pith.science/pith/CJEGRICZY52T4MQV7HOLIFR53R/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CJEGRICZY52T4MQV7HOLIFR53R/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:CJEGRICZY52T4MQV7HOLIFR53R","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":"303fc48d0b46a0673ef2148621d57d51ab9f6ec75b0f53f5003b8356e88ab9cd","cross_cats_sorted":["cs.CY","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-30T14:54:33Z","title_canon_sha256":"9e5991f17b9bc6e881745e84560e72e1117517aad133feddabb79ea2201adf77"},"schema_version":"1.0","source":{"id":"2607.28319","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.28319","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"arxiv_version","alias_value":"2607.28319v1","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.28319","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"pith_short_12","alias_value":"CJEGRICZY52T","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"pith_short_16","alias_value":"CJEGRICZY52T4MQV","created_at":"2026-07-31T01:37:13Z"},{"alias_kind":"pith_short_8","alias_value":"CJEGRICZ","created_at":"2026-07-31T01:37:13Z"}],"graph_snapshots":[{"event_id":"sha256:cacf39edc843a1ac93ec082edfaeec58d8849ec8db16108e709c230a1792eba7","target":"graph","created_at":"2026-07-31T01:37:13Z","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/2607.28319/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"This work presents Fairness Pruning, a lightweight structural intervention method designed for the management and future mitigation of demographic bias in large language models (LLMs). As a foundational empirical validation of this method, this work focuses on causal bias localization. Using minimally contrastive prompt pairs and inference-time activation capture, the method identifies neurons that react differentially when processing demographic attributes in GLU architectures, evaluating the signal at the down_proj input. Empirical evaluation was conducted on models of up to 3 billion parame","authors_text":"Alfonso Ure\\~na L\\'opez, Eugenio Mart\\'inez C\\'amara, Pere Martra","cross_cats":["cs.CY","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-30T14:54:33Z","title":"Fairness Pruning: Locating Demographic Bias in GLU-MLP Layers via Differential Activations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.28319","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":"sha256:0a7cedb4323d78869b7f0cc9c63841936db22f0aa88473668cdbc6a3f4589bd9","target":"record","created_at":"2026-07-31T01:37:13Z","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":"303fc48d0b46a0673ef2148621d57d51ab9f6ec75b0f53f5003b8356e88ab9cd","cross_cats_sorted":["cs.CY","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-07-30T14:54:33Z","title_canon_sha256":"9e5991f17b9bc6e881745e84560e72e1117517aad133feddabb79ea2201adf77"},"schema_version":"1.0","source":{"id":"2607.28319","kind":"arxiv","version":1}},"canonical_sha256":"124868a059c7753e3215f9dcb4163ddc7874fbcabf6a9fdaaca2fb94f4ddeb7a","receipt":{"builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"124868a059c7753e3215f9dcb4163ddc7874fbcabf6a9fdaaca2fb94f4ddeb7a","first_computed_at":"2026-07-31T01:37:13.756269Z","kind":"pith_receipt","last_reissued_at":"2026-07-31T01:37:13.756269Z","receipt_version":"0.3","signature_status":"unsigned_v0"},"source_id":"2607.28319","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0a7cedb4323d78869b7f0cc9c63841936db22f0aa88473668cdbc6a3f4589bd9","sha256:cacf39edc843a1ac93ec082edfaeec58d8849ec8db16108e709c230a1792eba7"],"state_sha256":"fef0f142ff0c16a9d4dc26e1bd5d98a52f0ab211416ab55c273e671f029e66bb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"72ovwZPjLSd4e08eHBH090WQgi1RJgITh+r9sthDzKEN09/d/UChHqdYRhclxPcb6kTPBDyCi/FN4aKouQ2dDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T02:42:31.431353Z","bundle_sha256":"28a54898bceeb116fc5c9405b975bdfe9d5c04ec0948cb1cdddcd02c1600c1bd"}}