Effective Proxy for Human Labeling: Ensemble Disagreement Scores in Large Language Models for Industrial NLP
This provides a cost-effective solution for industry practitioners needing to validate LLM performance in real-world settings, though it is incremental as it builds on existing ensemble methods.
The paper tackles the problem of expensive and delayed human labeling for assessing large language model (LLM) performance on unlabeled industrial data by proposing ensemble disagreement scores as a proxy, achieving a mean average error as low as 0.4% and outperforming silver labels by 13.8% on average in keyphrase extraction tasks.
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.