{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:D2PMXLZTVS5YAYHHJS6RTGU3RJ","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":"90f15a7fbc944fdfa408cbe6b145e55b409b078e064ac2408734a63f14a81414","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-16T13:44:06Z","title_canon_sha256":"41653e4a97fc34708a64fc4f9d2b0c4c6a70f0b64633c425c1dbb7189f8d59ad"},"schema_version":"1.0","source":{"id":"2605.16996","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.16996","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"arxiv_version","alias_value":"2605.16996v1","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16996","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_12","alias_value":"D2PMXLZTVS5Y","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_16","alias_value":"D2PMXLZTVS5YAYHH","created_at":"2026-05-20T00:03:35Z"},{"alias_kind":"pith_short_8","alias_value":"D2PMXLZT","created_at":"2026-05-20T00:03:35Z"}],"graph_snapshots":[{"event_id":"sha256:f5d55da3d54199c4863fa56526b02db1c82f755aca429536bbb9f7740b5ae45b","target":"graph","created_at":"2026-05-20T00:03:35Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Our results demonstrate that fine-tuning consistently reduces variance in questionnaire responses across five models, directly mitigating the evaluation fragility reported in pre-trained models. However, this newfound stability reveals a more fundamental limitation: accuracy on the full five-dimensional profile remains near chance, even when single-trait scores improve."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The IPIP-NEO questionnaire responses from LLMs validly and comprehensively measure the induced personality profile, and that the unguided essays contain sufficient trait-relevant cues to support faithful induction of the target five-dimensional profile."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Fine-tuning LLMs on essays reduces variance in IPIP-NEO responses across models but does not raise full five-trait profile accuracy above near-chance levels from unguided text."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Fine-tuning stabilizes LLM personality questionnaire scores but full-profile accuracy stays near chance."}],"snapshot_sha256":"a60b6640618cbedef44a9a55d1a94d12a4eb898ad9db8ddfabb27798975a88f4"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"ed57fe63a88412f21d67b18bda2db6e0af8deb990c25bca0ab02661362c1d5a3"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T21:01:19.050961Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T20:50:49.377930Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"citation_quote_validity","ran_at":"2026-05-19T19:49:57.560600Z","status":"skipped","version":"0.1.0"},{"findings_count":0,"name":"cited_work_retraction","ran_at":"2026-05-19T19:23:36.116463Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.202536Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.291934Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.16996/integrity.json","findings":[],"snapshot_sha256":"3f553111af421dbb59a5183c17b8c6c9e2917a692f49d0ca6111a77be026d30d","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Can large language models reliably express a human-like personality, or are they merely mimicking surface cues without a stable underlying profile? To investigate this, we induce personality in LLMs by fine-tuning them on the long-form essays, where each essay is associated with a target Big Five personality profile. We then evaluate the stability and fidelity of the induced personality using the IPIP-NEO questionnaire. Specifically, we ask: (i) does post-training (SFT, DPO, ORPO) stabilize questionnaire scores under prompt rephrasings, and (ii) can it induce target Big Five profiles from ungu","authors_text":"Iyiola E. Olatunji, Jacques Klein, Prateek Rajput, Tegawend\\'e F. Bissyand\\'e, Yewei Song","cross_cats":[],"headline":"Fine-tuning stabilizes LLM personality questionnaire scores but full-profile accuracy stays near chance.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-16T13:44:06Z","title":"Evaluation Drift in LLM Personality Induction: Are We Moving the Goalpost?"},"references":{"count":16,"internal_anchors":3,"resolved_work":16,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"InFindings of the Associa- tion for Computational Linguistics: EMNLP 2023, pages 2370–2386, Singapore","work_id":"26cf2f73-15a2-4856-a148-b4929d09762c","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Hans Christian, Derwin Suhartono, Andry Chowanda, and Kamal Z Zamli","work_id":"423d01ee-88a4-44d0-8336-3471bde8ab87","year":null},{"cited_arxiv_id":"2407.21783","doi":"","is_internal_anchor":true,"ref_index":3,"title":"The Llama 3 Herd of Models","work_id":"1549a635-88af-4ac1-acfe-51ae7bb53345","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Matej Gjurković and Jan Šnajder","work_id":"91f3b0b4-c23e-4a07-aeb0-5b3c98b33f0c","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"In2011 IEEE third international conference on privacy, security, risk and trust and 2011 IEEE third international conference on social computing, pages 149–156","work_id":"46248671-7c5f-40ad-b37b-b18746ecd5cd","year":2011}],"snapshot_sha256":"613212c98550714faef8b1d1e7ac3c071986de75adab7d1d7bf8cc35774b360b"},"source":{"id":"2605.16996","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T20:37:32.521866Z","id":"ad810bb2-9a0a-4c0c-b352-f170501a5933","model_set":{"reader":"grok-4.3"},"one_line_summary":"Fine-tuning LLMs on essays reduces variance in IPIP-NEO responses across models but does not raise full five-trait profile accuracy above near-chance levels from unguided text.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Fine-tuning stabilizes LLM personality questionnaire scores but full-profile accuracy stays near chance.","strongest_claim":"Our results demonstrate that fine-tuning consistently reduces variance in questionnaire responses across five models, directly mitigating the evaluation fragility reported in pre-trained models. However, this newfound stability reveals a more fundamental limitation: accuracy on the full five-dimensional profile remains near chance, even when single-trait scores improve.","weakest_assumption":"The IPIP-NEO questionnaire responses from LLMs validly and comprehensively measure the induced personality profile, and that the unguided essays contain sufficient trait-relevant cues to support faithful induction of the target five-dimensional profile."}},"verdict_id":"ad810bb2-9a0a-4c0c-b352-f170501a5933"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:f9affbce0503d902189caf37361f54109f0147dda4e5c5f95f9cfe5c41a34502","target":"record","created_at":"2026-05-20T00:03:35Z","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":"90f15a7fbc944fdfa408cbe6b145e55b409b078e064ac2408734a63f14a81414","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-16T13:44:06Z","title_canon_sha256":"41653e4a97fc34708a64fc4f9d2b0c4c6a70f0b64633c425c1dbb7189f8d59ad"},"schema_version":"1.0","source":{"id":"2605.16996","kind":"arxiv","version":1}},"canonical_sha256":"1e9ecbaf33acbb8060e74cbd199a9b8a7f929a4081f17701b25eb436ba9fccf7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1e9ecbaf33acbb8060e74cbd199a9b8a7f929a4081f17701b25eb436ba9fccf7","first_computed_at":"2026-05-20T00:03:35.087993Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:35.087993Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"GM4eLQxZemWIPN7l0ISdV6YoBkc7xVwxGju2gDat3VCYfddiCUjg1kYUp3ssOMT/fMcnb+qy0d+KUzRRnwVwCw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:35.088588Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.16996","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f9affbce0503d902189caf37361f54109f0147dda4e5c5f95f9cfe5c41a34502","sha256:f5d55da3d54199c4863fa56526b02db1c82f755aca429536bbb9f7740b5ae45b"],"state_sha256":"e1eedfe50f6ec1b624dc030242e7b89f431b06becdb657036053733df3cdf25c"}