{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MHLOUFP4UCNRVBLR64HGLRFOSU","short_pith_number":"pith:MHLOUFP4","schema_version":"1.0","canonical_sha256":"61d6ea15fca09b1a8571f70e65c4ae95052bb1fdbb7b3279980ed5ba62b337cc","source":{"kind":"arxiv","id":"2606.00467","version":1},"attestation_state":"computed","paper":{"title":"On the Limits of LLM Adaptability: Impact of Model-Internalized Priors on Annotation Task Performance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Etienne Casanova, Rafal Kocielnik, R. Michael Alvarez","submitted_at":"2026-05-30T01:21:14Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly used for zero-shot annotation and LLM-as-a-judge tasks, yet their reliability hinges on how model-internalized priors interact with user-provided instructions. We investigate three dimensions of this interaction: (1) how an LLM's familiarity with data and task definitions affects performance, (2) the extent to which additional information in prompts can correct zero-shot errors (\"decision stickiness\"), and (3) model susceptibility to misaligned task definitions. Through experiments on toxicity detection across diverse datasets (spanning social medi"},"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":"2606.00467","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-30T01:21:14Z","cross_cats_sorted":["cs.AI","cs.LG","stat.ML"],"title_canon_sha256":"53ce747ad67ea11c2f2d1548a1cbf24f77c2d54d05a391ff23cdb89501bf589b","abstract_canon_sha256":"42ed476264b92837fdf222c766ad5bd20fd7e51eeb46940caa537c69be8f47c3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:55.508881Z","signature_b64":"3EgjfNntaCYPBU+ABQbNpQqoLhkupqukMGqlf+eb1fExT/5nK+MUQLrBGcoP03E6ITM57qbQk57acalXr35JBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"61d6ea15fca09b1a8571f70e65c4ae95052bb1fdbb7b3279980ed5ba62b337cc","last_reissued_at":"2026-06-02T01:03:55.508539Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:55.508539Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Limits of LLM Adaptability: Impact of Model-Internalized Priors on Annotation Task Performance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Etienne Casanova, Rafal Kocielnik, R. Michael Alvarez","submitted_at":"2026-05-30T01:21:14Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly used for zero-shot annotation and LLM-as-a-judge tasks, yet their reliability hinges on how model-internalized priors interact with user-provided instructions. We investigate three dimensions of this interaction: (1) how an LLM's familiarity with data and task definitions affects performance, (2) the extent to which additional information in prompts can correct zero-shot errors (\"decision stickiness\"), and (3) model susceptibility to misaligned task definitions. Through experiments on toxicity detection across diverse datasets (spanning social medi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00467","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/2606.00467/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":"2606.00467","created_at":"2026-06-02T01:03:55.508587+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00467v1","created_at":"2026-06-02T01:03:55.508587+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00467","created_at":"2026-06-02T01:03:55.508587+00:00"},{"alias_kind":"pith_short_12","alias_value":"MHLOUFP4UCNR","created_at":"2026-06-02T01:03:55.508587+00:00"},{"alias_kind":"pith_short_16","alias_value":"MHLOUFP4UCNRVBLR","created_at":"2026-06-02T01:03:55.508587+00:00"},{"alias_kind":"pith_short_8","alias_value":"MHLOUFP4","created_at":"2026-06-02T01:03:55.508587+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/MHLOUFP4UCNRVBLR64HGLRFOSU","json":"https://pith.science/pith/MHLOUFP4UCNRVBLR64HGLRFOSU.json","graph_json":"https://pith.science/api/pith-number/MHLOUFP4UCNRVBLR64HGLRFOSU/graph.json","events_json":"https://pith.science/api/pith-number/MHLOUFP4UCNRVBLR64HGLRFOSU/events.json","paper":"https://pith.science/paper/MHLOUFP4"},"agent_actions":{"view_html":"https://pith.science/pith/MHLOUFP4UCNRVBLR64HGLRFOSU","download_json":"https://pith.science/pith/MHLOUFP4UCNRVBLR64HGLRFOSU.json","view_paper":"https://pith.science/paper/MHLOUFP4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00467&json=true","fetch_graph":"https://pith.science/api/pith-number/MHLOUFP4UCNRVBLR64HGLRFOSU/graph.json","fetch_events":"https://pith.science/api/pith-number/MHLOUFP4UCNRVBLR64HGLRFOSU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MHLOUFP4UCNRVBLR64HGLRFOSU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MHLOUFP4UCNRVBLR64HGLRFOSU/action/storage_attestation","attest_author":"https://pith.science/pith/MHLOUFP4UCNRVBLR64HGLRFOSU/action/author_attestation","sign_citation":"https://pith.science/pith/MHLOUFP4UCNRVBLR64HGLRFOSU/action/citation_signature","submit_replication":"https://pith.science/pith/MHLOUFP4UCNRVBLR64HGLRFOSU/action/replication_record"}},"created_at":"2026-06-02T01:03:55.508587+00:00","updated_at":"2026-06-02T01:03:55.508587+00:00"}