{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:I455A524BATPWTHXFIWSTKOIYM","short_pith_number":"pith:I455A524","schema_version":"1.0","canonical_sha256":"473bd0775c0826fb4cf72a2d29a9c8c3276379b297316671ea914352290e648c","source":{"kind":"arxiv","id":"2309.05619","version":2},"attestation_state":"computed","paper":{"title":"Effective Proxy for Human Labeling: Ensemble Disagreement Scores in Large Language Models for Industrial NLP","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aaron Colak, Daniel J Perry, Laksh Advani, Prashant Shiralkar, Wei Du, Yashmeet Gambhir, Zhengzheng Xing","submitted_at":"2023-09-11T17:07:01Z","abstract_excerpt":"Large language models (LLMs) have demonstrated significant capability to generalize across a large number of NLP tasks. For industry applications, it is imperative to assess the performance of the LLM on unlabeled production data from time to time to validate for a real-world setting. Human labeling to assess model error requires considerable expense and time delay. Here we demonstrate that ensemble disagreement scores work well as a proxy for human labeling for language models in zero-shot, few-shot, and fine-tuned settings, per our evaluation on keyphrase extraction (KPE) task. We measure fi"},"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":"2309.05619","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-09-11T17:07:01Z","cross_cats_sorted":[],"title_canon_sha256":"a95d70d5f3bd410fea2b4c95d28e8bd3d2d7e27306fe5ea973ba9ae2ab1e931c","abstract_canon_sha256":"4a90cfc210198436568aa03565fc8717a108dfd2c84551a652adbc09dfb6ffcc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:14:17.174975Z","signature_b64":"+1lKTwibXLaucTiLIt3hP5DCnlF4mTYocuSm7JEoUly6qcRDuXEbGvxv+8NEn4W8qSsY5AX5DclrlTb/luIHCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"473bd0775c0826fb4cf72a2d29a9c8c3276379b297316671ea914352290e648c","last_reissued_at":"2026-07-05T07:14:17.174533Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:14:17.174533Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Effective Proxy for Human Labeling: Ensemble Disagreement Scores in Large Language Models for Industrial NLP","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aaron Colak, Daniel J Perry, Laksh Advani, Prashant Shiralkar, Wei Du, Yashmeet Gambhir, Zhengzheng Xing","submitted_at":"2023-09-11T17:07:01Z","abstract_excerpt":"Large language models (LLMs) have demonstrated significant capability to generalize across a large number of NLP tasks. For industry applications, it is imperative to assess the performance of the LLM on unlabeled production data from time to time to validate for a real-world setting. Human labeling to assess model error requires considerable expense and time delay. Here we demonstrate that ensemble disagreement scores work well as a proxy for human labeling for language models in zero-shot, few-shot, and fine-tuned settings, per our evaluation on keyphrase extraction (KPE) task. We measure fi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.05619","kind":"arxiv","version":2},"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/2309.05619/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":"2309.05619","created_at":"2026-07-05T07:14:17.174590+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.05619v2","created_at":"2026-07-05T07:14:17.174590+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.05619","created_at":"2026-07-05T07:14:17.174590+00:00"},{"alias_kind":"pith_short_12","alias_value":"I455A524BATP","created_at":"2026-07-05T07:14:17.174590+00:00"},{"alias_kind":"pith_short_16","alias_value":"I455A524BATPWTHX","created_at":"2026-07-05T07:14:17.174590+00:00"},{"alias_kind":"pith_short_8","alias_value":"I455A524","created_at":"2026-07-05T07:14:17.174590+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/I455A524BATPWTHXFIWSTKOIYM","json":"https://pith.science/pith/I455A524BATPWTHXFIWSTKOIYM.json","graph_json":"https://pith.science/api/pith-number/I455A524BATPWTHXFIWSTKOIYM/graph.json","events_json":"https://pith.science/api/pith-number/I455A524BATPWTHXFIWSTKOIYM/events.json","paper":"https://pith.science/paper/I455A524"},"agent_actions":{"view_html":"https://pith.science/pith/I455A524BATPWTHXFIWSTKOIYM","download_json":"https://pith.science/pith/I455A524BATPWTHXFIWSTKOIYM.json","view_paper":"https://pith.science/paper/I455A524","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.05619&json=true","fetch_graph":"https://pith.science/api/pith-number/I455A524BATPWTHXFIWSTKOIYM/graph.json","fetch_events":"https://pith.science/api/pith-number/I455A524BATPWTHXFIWSTKOIYM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I455A524BATPWTHXFIWSTKOIYM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I455A524BATPWTHXFIWSTKOIYM/action/storage_attestation","attest_author":"https://pith.science/pith/I455A524BATPWTHXFIWSTKOIYM/action/author_attestation","sign_citation":"https://pith.science/pith/I455A524BATPWTHXFIWSTKOIYM/action/citation_signature","submit_replication":"https://pith.science/pith/I455A524BATPWTHXFIWSTKOIYM/action/replication_record"}},"created_at":"2026-07-05T07:14:17.174590+00:00","updated_at":"2026-07-05T07:14:17.174590+00:00"}