{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:LPGJKIMKUAC6GQNCEDRRO6PHP5","short_pith_number":"pith:LPGJKIMK","canonical_record":{"source":{"id":"2506.00253","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T21:41:44Z","cross_cats_sorted":["cs.AI","cs.CY"],"title_canon_sha256":"96e3aa5ced82ab1cc12657871f99a12a2e0946ba911cb3ae2c8f78198140c7f5","abstract_canon_sha256":"a2ec857eaecb33329bb44c6c8d0b94579504435af75924b977368f9416df67fe"},"schema_version":"1.0"},"canonical_sha256":"5bcc95218aa005e341a220e31779e77f7f7cdd0c95e7ae1e54a14dadd40f55f4","source":{"kind":"arxiv","id":"2506.00253","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.00253","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"arxiv_version","alias_value":"2506.00253v3","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.00253","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"pith_short_12","alias_value":"LPGJKIMKUAC6","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"pith_short_16","alias_value":"LPGJKIMKUAC6GQNC","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"pith_short_8","alias_value":"LPGJKIMK","created_at":"2026-07-05T11:18:22Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:LPGJKIMKUAC6GQNCEDRRO6PHP5","target":"record","payload":{"canonical_record":{"source":{"id":"2506.00253","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T21:41:44Z","cross_cats_sorted":["cs.AI","cs.CY"],"title_canon_sha256":"96e3aa5ced82ab1cc12657871f99a12a2e0946ba911cb3ae2c8f78198140c7f5","abstract_canon_sha256":"a2ec857eaecb33329bb44c6c8d0b94579504435af75924b977368f9416df67fe"},"schema_version":"1.0"},"canonical_sha256":"5bcc95218aa005e341a220e31779e77f7f7cdd0c95e7ae1e54a14dadd40f55f4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:18:22.649340Z","signature_b64":"C+7HbfAy0rkZ3PlSHXQXEVzIJAYHkroHKKaOElRCYS0EuiPzGUGtdIVCSGpZLF0IRaFuow/zmpd89fPKnNcWAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5bcc95218aa005e341a220e31779e77f7f7cdd0c95e7ae1e54a14dadd40f55f4","last_reissued_at":"2026-07-05T11:18:22.648750Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:18:22.648750Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.00253","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:18:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RxHekgZ343UWSM97IIcUM8gS2Sm88BHxbGhd8jG2K/IGN8UQnD2yed94y69NnXKeLncS2qxIcx5gErdA5flGAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T01:48:56.242818Z"},"content_sha256":"44a40ef0bacf8ec9ead4710d2d6c3e191a427387df75548058242aa65cb42b8f","schema_version":"1.0","event_id":"sha256:44a40ef0bacf8ec9ead4710d2d6c3e191a427387df75548058242aa65cb42b8f"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:LPGJKIMKUAC6GQNCEDRRO6PHP5","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Aligned but Blind: Alignment Increases Implicit Bias by Reducing Awareness of Race","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CY"],"primary_cat":"cs.CL","authors_text":"Chengzhi Mao, Lihao Sun, Valentin Hofmann, Xuechunzi Bai","submitted_at":"2025-05-30T21:41:44Z","abstract_excerpt":"Although value-aligned language models (LMs) appear unbiased in explicit bias evaluations, they often exhibit stereotypes in implicit word association tasks, raising concerns about their fair usage. We investigate the mechanisms behind this discrepancy and find that alignment surprisingly amplifies implicit bias in model outputs. Specifically, we show that aligned LMs, unlike their unaligned counterparts, overlook racial concepts in early internal representations when the context is ambiguous. Not representing race likely fails to activate safety guardrails, leading to unintended biases. Inspi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.00253","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/2506.00253/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":"2026-07-05T11:18:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XQXp8HUAzT7wgL3B95qh3wanhnq9kqXuW3L0n+Amu0SviDwLAh5KlbeG/DMYnPyFUFPyKIK8l9f+isGAF/VQCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T01:48:56.243720Z"},"content_sha256":"bb7bbaa0fb4f0ba6b46a0b579e8f5d2e829c4e2c3f2e772723a09a5ce4271bf2","schema_version":"1.0","event_id":"sha256:bb7bbaa0fb4f0ba6b46a0b579e8f5d2e829c4e2c3f2e772723a09a5ce4271bf2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LPGJKIMKUAC6GQNCEDRRO6PHP5/bundle.json","state_url":"https://pith.science/pith/LPGJKIMKUAC6GQNCEDRRO6PHP5/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LPGJKIMKUAC6GQNCEDRRO6PHP5/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":"2026-08-10T01:48:56Z","links":{"resolver":"https://pith.science/pith/LPGJKIMKUAC6GQNCEDRRO6PHP5","bundle":"https://pith.science/pith/LPGJKIMKUAC6GQNCEDRRO6PHP5/bundle.json","state":"https://pith.science/pith/LPGJKIMKUAC6GQNCEDRRO6PHP5/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LPGJKIMKUAC6GQNCEDRRO6PHP5/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:LPGJKIMKUAC6GQNCEDRRO6PHP5","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":"a2ec857eaecb33329bb44c6c8d0b94579504435af75924b977368f9416df67fe","cross_cats_sorted":["cs.AI","cs.CY"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T21:41:44Z","title_canon_sha256":"96e3aa5ced82ab1cc12657871f99a12a2e0946ba911cb3ae2c8f78198140c7f5"},"schema_version":"1.0","source":{"id":"2506.00253","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.00253","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"arxiv_version","alias_value":"2506.00253v3","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.00253","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"pith_short_12","alias_value":"LPGJKIMKUAC6","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"pith_short_16","alias_value":"LPGJKIMKUAC6GQNC","created_at":"2026-07-05T11:18:22Z"},{"alias_kind":"pith_short_8","alias_value":"LPGJKIMK","created_at":"2026-07-05T11:18:22Z"}],"graph_snapshots":[{"event_id":"sha256:bb7bbaa0fb4f0ba6b46a0b579e8f5d2e829c4e2c3f2e772723a09a5ce4271bf2","target":"graph","created_at":"2026-07-05T11:18:22Z","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/2506.00253/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Although value-aligned language models (LMs) appear unbiased in explicit bias evaluations, they often exhibit stereotypes in implicit word association tasks, raising concerns about their fair usage. We investigate the mechanisms behind this discrepancy and find that alignment surprisingly amplifies implicit bias in model outputs. Specifically, we show that aligned LMs, unlike their unaligned counterparts, overlook racial concepts in early internal representations when the context is ambiguous. Not representing race likely fails to activate safety guardrails, leading to unintended biases. Inspi","authors_text":"Chengzhi Mao, Lihao Sun, Valentin Hofmann, Xuechunzi Bai","cross_cats":["cs.AI","cs.CY"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T21:41:44Z","title":"Aligned but Blind: Alignment Increases Implicit Bias by Reducing Awareness of Race"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.00253","kind":"arxiv","version":3},"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":"sha256:44a40ef0bacf8ec9ead4710d2d6c3e191a427387df75548058242aa65cb42b8f","target":"record","created_at":"2026-07-05T11:18:22Z","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":"a2ec857eaecb33329bb44c6c8d0b94579504435af75924b977368f9416df67fe","cross_cats_sorted":["cs.AI","cs.CY"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-30T21:41:44Z","title_canon_sha256":"96e3aa5ced82ab1cc12657871f99a12a2e0946ba911cb3ae2c8f78198140c7f5"},"schema_version":"1.0","source":{"id":"2506.00253","kind":"arxiv","version":3}},"canonical_sha256":"5bcc95218aa005e341a220e31779e77f7f7cdd0c95e7ae1e54a14dadd40f55f4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5bcc95218aa005e341a220e31779e77f7f7cdd0c95e7ae1e54a14dadd40f55f4","first_computed_at":"2026-07-05T11:18:22.648750Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:18:22.648750Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"C+7HbfAy0rkZ3PlSHXQXEVzIJAYHkroHKKaOElRCYS0EuiPzGUGtdIVCSGpZLF0IRaFuow/zmpd89fPKnNcWAw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:18:22.649340Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.00253","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:44a40ef0bacf8ec9ead4710d2d6c3e191a427387df75548058242aa65cb42b8f","sha256:bb7bbaa0fb4f0ba6b46a0b579e8f5d2e829c4e2c3f2e772723a09a5ce4271bf2"],"state_sha256":"7e8a1914bb0eaaded1538ef781cf62924e234dda3476960e590b4a38f0b8e7d9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iH7SyNuL0gkyKCFTpQg92MjZeW6VloY1s4x9UMkVMtJ3+YuEXJsfwS0pFLOHq/MyTz2++dNnh6S5NyVuTJCPCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T01:48:56.251746Z","bundle_sha256":"397ab47ec7b95205575d1d77b98bc0066814c56a4eabe66409bbc888ed0f620c"}}