Effective Proxy for Human Labeling: Ensemble Disagreement Scores in Large Language Models for Industrial NLP
Abstract
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 fidelity of the results by comparing to true error measured from human labeled ground truth. We contrast with the alternative of using another LLM as a source of machine labels, or silver labels. Results across various languages and domains show disagreement scores provide a better estimation of model performance with mean average error (MAE) as low as 0.4% and on average 13.8% better than using silver labels.
Keywords
Cite
@article{arxiv.2309.05619,
title = {Effective Proxy for Human Labeling: Ensemble Disagreement Scores in Large Language Models for Industrial NLP},
author = {Wei Du and Laksh Advani and Yashmeet Gambhir and Daniel J Perry and Prashant Shiralkar and Zhengzheng Xing and Aaron Colak},
journal= {arXiv preprint arXiv:2309.05619},
year = {2023}
}
Comments
Camera ready version for 2023 EMNLP (The Third Workshop on Natural Language Generation, Evaluation, and Metrics (GEM))