{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:SKDPG5DCEOZZKMDCJC3XGLEZSG","short_pith_number":"pith:SKDPG5DC","canonical_record":{"source":{"id":"2509.04373","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T16:32:18Z","cross_cats_sorted":[],"title_canon_sha256":"6df7a511bb5667b0cb8e54580a5d665ebd2ffbd063ed11106532eb308f8beed2","abstract_canon_sha256":"111bf55a7e1fef3355ff3b917c5ad0e85b0c48ec4f374a38efd81b23c838f448"},"schema_version":"1.0"},"canonical_sha256":"9286f3746223b395306248b7732c9991996d23c4582593637a02d0d19c32ff2f","source":{"kind":"arxiv","id":"2509.04373","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.04373","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"arxiv_version","alias_value":"2509.04373v3","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.04373","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"pith_short_12","alias_value":"SKDPG5DCEOZZ","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"pith_short_16","alias_value":"SKDPG5DCEOZZKMDC","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"pith_short_8","alias_value":"SKDPG5DC","created_at":"2026-07-05T12:08:07Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:SKDPG5DCEOZZKMDCJC3XGLEZSG","target":"record","payload":{"canonical_record":{"source":{"id":"2509.04373","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T16:32:18Z","cross_cats_sorted":[],"title_canon_sha256":"6df7a511bb5667b0cb8e54580a5d665ebd2ffbd063ed11106532eb308f8beed2","abstract_canon_sha256":"111bf55a7e1fef3355ff3b917c5ad0e85b0c48ec4f374a38efd81b23c838f448"},"schema_version":"1.0"},"canonical_sha256":"9286f3746223b395306248b7732c9991996d23c4582593637a02d0d19c32ff2f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:08:07.539790Z","signature_b64":"efT2zmaLOJ+kfBfIqVIEDyQPhTqSQXzQWrknL589oozsjryrYoUEXwriV9+iPkbvHEKbOqq3e7AT6S3n3ittDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9286f3746223b395306248b7732c9991996d23c4582593637a02d0d19c32ff2f","last_reissued_at":"2026-07-05T12:08:07.539109Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:08:07.539109Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2509.04373","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":"2026-07-05T12:08:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"woAKBiLXR+XstZwnzrEoWeBTTbwdWKeBJ9ySSQjsIbJNNkmO9I+DXXlqC8glk2/S8s/h3PNT+Ib82PN1K79eCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T19:46:24.800465Z"},"content_sha256":"d382600040b499b909b2f5b030ec3e31390a9a77155ead4d0d4ea0d96352dafa","schema_version":"1.0","event_id":"sha256:d382600040b499b909b2f5b030ec3e31390a9a77155ead4d0d4ea0d96352dafa"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:SKDPG5DCEOZZKMDCJC3XGLEZSG","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Measuring Bias or Measuring the Task: Understanding the Brittle Nature of LLM Gender Biases","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bufan Gao, Elisa Kreiss","submitted_at":"2025-09-04T16:32:18Z","abstract_excerpt":"As LLMs are increasingly applied in socially impactful settings, concerns about gender bias have prompted growing efforts both to measure and mitigate such bias. These efforts often rely on evaluation tasks that differ from natural language distributions, as they typically involve carefully constructed task prompts that overtly or covertly signal the presence of gender bias-related content. In this paper, we examine how signaling the evaluative purpose of a task impacts measured gender bias in LLMs. Concretely, we test models under prompt conditions that (1) make the testing context salient, a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.04373","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/2509.04373/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":"2026-07-05T12:08:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FA3HroJ+mimRcbL+irSdnPdfyozI/6PLl3jyDR1Woj0g2LNQ7F0HtrCfBKHla6qt2lTwFE05WizIQvXZqgIyCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T19:46:24.800980Z"},"content_sha256":"5c8b68e3b3268c78798eab60274c8c61c5ab6091d3172c1f4892b6fade19601b","schema_version":"1.0","event_id":"sha256:5c8b68e3b3268c78798eab60274c8c61c5ab6091d3172c1f4892b6fade19601b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SKDPG5DCEOZZKMDCJC3XGLEZSG/bundle.json","state_url":"https://pith.science/pith/SKDPG5DCEOZZKMDCJC3XGLEZSG/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SKDPG5DCEOZZKMDCJC3XGLEZSG/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":"2026-08-05T19:46:24Z","links":{"resolver":"https://pith.science/pith/SKDPG5DCEOZZKMDCJC3XGLEZSG","bundle":"https://pith.science/pith/SKDPG5DCEOZZKMDCJC3XGLEZSG/bundle.json","state":"https://pith.science/pith/SKDPG5DCEOZZKMDCJC3XGLEZSG/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SKDPG5DCEOZZKMDCJC3XGLEZSG/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:SKDPG5DCEOZZKMDCJC3XGLEZSG","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":"111bf55a7e1fef3355ff3b917c5ad0e85b0c48ec4f374a38efd81b23c838f448","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T16:32:18Z","title_canon_sha256":"6df7a511bb5667b0cb8e54580a5d665ebd2ffbd063ed11106532eb308f8beed2"},"schema_version":"1.0","source":{"id":"2509.04373","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.04373","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"arxiv_version","alias_value":"2509.04373v3","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.04373","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"pith_short_12","alias_value":"SKDPG5DCEOZZ","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"pith_short_16","alias_value":"SKDPG5DCEOZZKMDC","created_at":"2026-07-05T12:08:07Z"},{"alias_kind":"pith_short_8","alias_value":"SKDPG5DC","created_at":"2026-07-05T12:08:07Z"}],"graph_snapshots":[{"event_id":"sha256:5c8b68e3b3268c78798eab60274c8c61c5ab6091d3172c1f4892b6fade19601b","target":"graph","created_at":"2026-07-05T12:08:07Z","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/2509.04373/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As LLMs are increasingly applied in socially impactful settings, concerns about gender bias have prompted growing efforts both to measure and mitigate such bias. These efforts often rely on evaluation tasks that differ from natural language distributions, as they typically involve carefully constructed task prompts that overtly or covertly signal the presence of gender bias-related content. In this paper, we examine how signaling the evaluative purpose of a task impacts measured gender bias in LLMs. Concretely, we test models under prompt conditions that (1) make the testing context salient, a","authors_text":"Bufan Gao, Elisa Kreiss","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T16:32:18Z","title":"Measuring Bias or Measuring the Task: Understanding the Brittle Nature of LLM Gender Biases"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.04373","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:d382600040b499b909b2f5b030ec3e31390a9a77155ead4d0d4ea0d96352dafa","target":"record","created_at":"2026-07-05T12:08:07Z","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":"111bf55a7e1fef3355ff3b917c5ad0e85b0c48ec4f374a38efd81b23c838f448","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T16:32:18Z","title_canon_sha256":"6df7a511bb5667b0cb8e54580a5d665ebd2ffbd063ed11106532eb308f8beed2"},"schema_version":"1.0","source":{"id":"2509.04373","kind":"arxiv","version":3}},"canonical_sha256":"9286f3746223b395306248b7732c9991996d23c4582593637a02d0d19c32ff2f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9286f3746223b395306248b7732c9991996d23c4582593637a02d0d19c32ff2f","first_computed_at":"2026-07-05T12:08:07.539109Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T12:08:07.539109Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"efT2zmaLOJ+kfBfIqVIEDyQPhTqSQXzQWrknL589oozsjryrYoUEXwriV9+iPkbvHEKbOqq3e7AT6S3n3ittDg==","signature_status":"signed_v1","signed_at":"2026-07-05T12:08:07.539790Z","signed_message":"canonical_sha256_bytes"},"source_id":"2509.04373","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d382600040b499b909b2f5b030ec3e31390a9a77155ead4d0d4ea0d96352dafa","sha256:5c8b68e3b3268c78798eab60274c8c61c5ab6091d3172c1f4892b6fade19601b"],"state_sha256":"c93f2785a8e43fa22e5512538bf961fcdb0b7aef65f22fbcd13900274aa13c15"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CxBPjUuJJ86NAM0GgiXwPyTrvV/eP+kY2Y0g7FO34qG0GleclMJMhOjg8TSvkCYXrNjCtvyqzjIOMo01n1DqAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T19:46:24.806590Z","bundle_sha256":"810a19fbb93da07ec6c166d9875bfabc8c75de947055d1af651846cbd64686f3"}}