What to Ask Next? Probing the Imaginative Reasoning of LLMs with TurtleSoup Puzzles

TL;DR

This paper introduces TurtleSoup-Bench, a bilingual interactive benchmark, and Mosaic-Agent, a novel evaluation agent, to assess the imaginative reasoning capabilities of Large Language Models through dynamic, exploratory puzzles.

Contribution

It presents the first large-scale, bilingual benchmark and an agent specifically designed to evaluate LLMs' imaginative reasoning in a dynamic, hypothesis-driven environment.

Findings

LLMs show significant limitations in imaginative reasoning.

Common failure patterns identified in LLMs' reasoning processes.

Performance gap between LLMs and human reasoning capabilities.

Abstract

We investigate the capacity of Large Language Models (LLMs) for imaginative reasoning--the proactive construction, testing, and revision of hypotheses in information-sparse environments. Existing benchmarks, often static or focused on social deduction, fail to capture the dynamic, exploratory nature of this reasoning process. To address this gap, we introduce a comprehensive research framework based on the classic "Turtle Soup" game, integrating a benchmark, an agent, and an evaluation protocol. We present TurtleSoup-Bench, the first large-scale, bilingual, interactive benchmark for imaginative reasoning, comprising 800 turtle soup puzzles sourced from both the Internet and expert authors. We also propose Mosaic-Agent, a novel agent designed to assess LLMs' performance in this setting. To evaluate reasoning quality, we develop a multi-dimensional protocol measuring logical consistency, detail completion, and conclusion alignment. Experiments with leading LLMs reveal clear capability limits, common failure patterns, and a significant performance gap compared to humans. Our work offers new insights into LLMs' imaginative reasoning and establishes a foundation for future research on exploratory agent behavior.