The Ability of Large Language Models to Evaluate Constraint-satisfaction in Agent Responses to Open-ended Requests

TL;DR

This paper investigates the ability of large language models to evaluate whether agent responses satisfy complex constraints in open-ended requests, introducing a new dataset and benchmarking their reasoning and arithmetic skills.

Contribution

The paper presents the ACS dataset for evaluating constraint satisfaction and benchmarks LLMs, revealing their limitations and the challenges of few-shot prompting in this context.

Findings

Most LLMs show significant room for improvement in constraint evaluation.

Errors mainly arise from reasoning difficulties in the models.

Few-shot prompting often degrades model performance.

Abstract

Generative AI agents are often expected to respond to complex user requests that have No One Right Answer (NORA), e.g., "design a vegetarian meal plan below 1800 calories". Such requests may entail a set of constraints that the agent should adhere to. To successfully develop agents for NORA scenarios, an accurate automatic evaluation framework is essential, and specifically - one capable of validating the satisfaction of constraints in the agent's response. Recently, large language models (LLMs) have been adopted as versatile evaluators for many NORA tasks, but their ability to evaluate constraint-satisfaction in generated text remains unclear. To study this, we develop and release a novel Arithmetic Constraint-Satisfaction (ACS) benchmarking dataset. The dataset consists of complex user requests with corresponding constraints, agent responses and human labels indicating each constraint's satisfaction level in the response. A unique property of this dataset is that validating many of its constraints requires reviewing the response as a whole (in contrast to many other benchmarks that require the validation of a single independent item). Moreover, it assesses LLMs in performing reasoning, in-context data extraction, arithmetic calculations, and counting. We then benchmark both open and proprietary LLMs on evaluating constraint-satisfaction, and show that most models still have a significant headroom for improvement, and that errors primarily stem from reasoning issues. In addition, most models exhibit a skewed constraint-satisfaction prediction pattern, with higher accuracy where the ground-truth label is "satisfied". Lastly, few-shot prompting for our task proved to be rather challenging, since many of the studied models showed a degradation in performance when it was introduced.