Evaluating LLM Understanding via Structured Tabular Decision Simulations

TL;DR

This paper introduces Structured Tabular Decision Simulations (STaDS), a new evaluation framework for assessing whether large language models truly understand decision factors across diverse domains, beyond mere accuracy.

Contribution

The paper presents STaDS, a novel suite of decision-based evaluation settings that measure LLM understanding through decision factor reliance and comprehension across multiple domains.

Findings

Most models struggle with consistent accuracy across domains

Models can be accurate but rely on incorrect decision factors

Frequent mismatches between rationales and actual decision drivers

Abstract

Large language models (LLMs) often achieve impressive predictive accuracy, yet correctness alone does not imply genuine understanding. True LLM understanding, analogous to human expertise, requires making consistent, well-founded decisions across multiple instances and diverse domains, relying on relevant and domain-grounded decision factors. We introduce Structured Tabular Decision Simulations (STaDS), a suite of expert-like decision settings that evaluate LLMs as if they were professionals undertaking structured decision ``exams''. In this context, understanding is defined as the ability to identify and rely on the correct decision factors, features that determine outcomes within a domain. STaDS jointly assesses understanding through: (i) question and instruction comprehension, (ii) knowledge-based prediction, and (iii) reliance on relevant decision factors. By analyzing 9 frontier LLMs across 15 diverse decision settings, we find that (a) most models struggle to achieve consistently strong accuracy across diverse domains; (b) models can be accurate yet globally unfaithful, and there are frequent mismatches between stated rationales and factors driving predictions. Our findings highlight the need for global-level understanding evaluation protocols and advocate for novel frameworks that go beyond accuracy to enhance LLMs' understanding ability.