TurnaboutLLM: A Deductive Reasoning Benchmark from Detective Games

TL;DR

TurnaboutLLM is a new benchmark using detective game narratives to evaluate and analyze the deductive reasoning capabilities of large language models, revealing current limitations and influencing factors.

Contribution

The paper introduces a novel deductive reasoning benchmark based on detective games, providing a new dataset and framework for assessing LLMs in complex narrative contexts.

Findings

LLMs show limitations in deductive reasoning within narrative environments

Chain-of-Thought prompting has limited effectiveness on this benchmark

Context size and reasoning steps significantly impact model performance

Abstract

This paper introduces TurnaboutLLM, a novel framework and dataset for evaluating the deductive reasoning abilities of Large Language Models (LLMs) by leveraging the interactive gameplay of detective games Ace Attorney and Danganronpa. The framework tasks LLMs with identifying contradictions between testimonies and evidences within long narrative contexts, a challenging task due to the large answer space and diverse reasoning types presented by its questions. We evaluate twelve state-of-the-art LLMs on the dataset, hinting at limitations of popular strategies for enhancing deductive reasoning such as extensive thinking and Chain-of-Thought prompting. The results also suggest varying effects of context size, the number of reasoning step and answer space size on model performance. Overall, TurnaboutLLM presents a substantial challenge for LLMs' deductive reasoning abilities in complex, narrative-rich environments.