Bayesian Statistical Modeling with Predictors from LLMs

TL;DR

This paper evaluates the human-likeness of LLM predictions in decision tasks using Bayesian models, revealing that LLMs do not capture individual variance but can approximate aggregate behavior with specific methods.

Contribution

It introduces Bayesian statistical modeling to assess LLMs' alignment with human data and explores methods to derive meaningful distributional predictions from LLMs.

Findings

LLMs do not capture variance at the individual item level.

Some methods of deriving condition-level predictions fit human data adequately.

Assessment of LLM performance depends on methodological choices.

Abstract

State of the art large language models (LLMs) have shown impressive performance on a variety of benchmark tasks and are increasingly used as components in larger applications, where LLM-based predictions serve as proxies for human judgements or decision. This raises questions about the human-likeness of LLM-derived information, alignment with human intuition, and whether LLMs could possibly be considered (parts of) explanatory models of (aspects of) human cognition or language use. To shed more light on these issues, we here investigate the human-likeness of LLMs' predictions for multiple-choice decision tasks from the perspective of Bayesian statistical modeling. Using human data from a forced-choice experiment on pragmatic language use, we find that LLMs do not capture the variance in the human data at the item-level. We suggest different ways of deriving full distributional predictions from LLMs for aggregate, condition-level data, and find that some, but not all ways of obtaining condition-level predictions yield adequate fits to human data. These results suggests that assessment of LLM performance depends strongly on seemingly subtle choices in methodology, and that LLMs are at best predictors of human behavior at the aggregate, condition-level, for which they are, however, not designed to, or usually used to, make predictions in the first place.