ToMChallenges: A Principle-Guided Dataset and Diverse Evaluation Tasks for Exploring Theory of Mind
This work addresses the inconsistent evaluation of ToM in LLMs, which is crucial for AI applications requiring social understanding, but it is incremental as it builds on existing tests and datasets.
The paper tackles the problem of evaluating Theory of Mind (ToM) in large language models (LLMs) by introducing ToMChallenges, a dataset based on Sally-Anne and Smarties tests with diverse tasks, and testing models like davinci, turbo, and gpt-4, finding that LLMs show inconsistent behaviors across prompts and tasks, making robust ToM performance a challenge.
Theory of Mind (ToM), the capacity to comprehend the mental states of distinct individuals, is essential for numerous practical applications. With the development of large language models (LLMs), there is a heated debate about whether they are able to perform ToM tasks. Previous studies have used different tasks and prompts to test the ToM on LLMs and the results are inconsistent: some studies asserted these models are capable of exhibiting ToM, while others suggest the opposite. In this study, We present ToMChallenges, a dataset for comprehensively evaluating the Theory of Mind based on the Sally-Anne and Smarties tests with a diverse set of tasks. In addition, we also propose an auto-grader to streamline the answer evaluation process. We tested three models: davinci, turbo, and gpt-4. Our evaluation results and error analyses show that LLMs have inconsistent behaviors across prompts and tasks. Performing the ToM tasks robustly remains a challenge for the LLMs. In addition, our paper wants to raise awareness in evaluating the ToM in LLMs and we want to invite more discussion on how to design the prompts and tasks for ToM tasks that can better assess the LLMs' ability.