{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NCD52TRXMF35DNUJCP4E7ALTTX","short_pith_number":"pith:NCD52TRX","schema_version":"1.0","canonical_sha256":"6887dd4e376177d1b68913f84f81739dc68b8a7d9b9bc374a7bc60881dd1b3d5","source":{"kind":"arxiv","id":"2606.12426","version":1},"attestation_state":"computed","paper":{"title":"Two Wrongs, No Right: Auditing Social-Desirability Bias in LLM Annotators for Computational Social Science","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CY","authors_text":"Varun Kotte","submitted_at":"2026-05-12T08:14:10Z","abstract_excerpt":"LLM annotators are increasingly used in computational social science (CSS), but it is unclear whether their alignment-shaped errors preserve the empirical conclusions a researcher would report. We audit three open-source 7B instruction-tuned models (Zephyr, Mistral-Instruct, Qwen2.5-Instruct) across six TweetEval tasks under four prompt conditions (72 cells) and find that social-desirability failures do not run in a single direction. Zephyr exhibits leniency bias, systematically under-applying harmful labels (offensive language: false benign rate 0.729, false alarm rate 0.031). Mistral and Qwe"},"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":"2606.12426","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CY","submitted_at":"2026-05-12T08:14:10Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"afeb842d3892c066048b8545f65394dd969e967e42fdabc54e60bf33e2447155","abstract_canon_sha256":"107a3ba0fe3f35bc8a5c74927f56928da5de4de6e38ed4957f6b6e1e9366694c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-12T00:07:49.857342Z","signature_b64":"Y9w1lKtnAHlrjaB/JyBFUXKaTRlPD13S2G83/Z81WctdixeFxT1d35uVRHruuFw6Ezjllm3tXZ8KV+ligi3eBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6887dd4e376177d1b68913f84f81739dc68b8a7d9b9bc374a7bc60881dd1b3d5","last_reissued_at":"2026-06-12T00:07:49.856397Z","signature_status":"signed_v1","first_computed_at":"2026-06-12T00:07:49.856397Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Two Wrongs, No Right: Auditing Social-Desirability Bias in LLM Annotators for Computational Social Science","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CY","authors_text":"Varun Kotte","submitted_at":"2026-05-12T08:14:10Z","abstract_excerpt":"LLM annotators are increasingly used in computational social science (CSS), but it is unclear whether their alignment-shaped errors preserve the empirical conclusions a researcher would report. We audit three open-source 7B instruction-tuned models (Zephyr, Mistral-Instruct, Qwen2.5-Instruct) across six TweetEval tasks under four prompt conditions (72 cells) and find that social-desirability failures do not run in a single direction. Zephyr exhibits leniency bias, systematically under-applying harmful labels (offensive language: false benign rate 0.729, false alarm rate 0.031). Mistral and Qwe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.12426","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/2606.12426/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":"2606.12426","created_at":"2026-06-12T00:07:49.856548+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.12426v1","created_at":"2026-06-12T00:07:49.856548+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.12426","created_at":"2026-06-12T00:07:49.856548+00:00"},{"alias_kind":"pith_short_12","alias_value":"NCD52TRXMF35","created_at":"2026-06-12T00:07:49.856548+00:00"},{"alias_kind":"pith_short_16","alias_value":"NCD52TRXMF35DNUJ","created_at":"2026-06-12T00:07:49.856548+00:00"},{"alias_kind":"pith_short_8","alias_value":"NCD52TRX","created_at":"2026-06-12T00:07:49.856548+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX","json":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX.json","graph_json":"https://pith.science/api/pith-number/NCD52TRXMF35DNUJCP4E7ALTTX/graph.json","events_json":"https://pith.science/api/pith-number/NCD52TRXMF35DNUJCP4E7ALTTX/events.json","paper":"https://pith.science/paper/NCD52TRX"},"agent_actions":{"view_html":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX","download_json":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX.json","view_paper":"https://pith.science/paper/NCD52TRX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.12426&json=true","fetch_graph":"https://pith.science/api/pith-number/NCD52TRXMF35DNUJCP4E7ALTTX/graph.json","fetch_events":"https://pith.science/api/pith-number/NCD52TRXMF35DNUJCP4E7ALTTX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX/action/storage_attestation","attest_author":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX/action/author_attestation","sign_citation":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX/action/citation_signature","submit_replication":"https://pith.science/pith/NCD52TRXMF35DNUJCP4E7ALTTX/action/replication_record"}},"created_at":"2026-06-12T00:07:49.856548+00:00","updated_at":"2026-06-12T00:07:49.856548+00:00"}