WILT: A Multi-Turn, Memorization-Robust Inductive Logic Benchmark for LLMs

TL;DR

WILT is a multi-turn reasoning benchmark designed to evaluate LLMs' ability to infer hidden logical functions through interactive testing, exposing their limitations in complex reasoning tasks beyond memorization.

Contribution

This paper introduces WILT, a novel multi-turn reasoning benchmark that challenges LLMs to infer hidden functions through interactive testing, resisting memorization and highlighting reasoning weaknesses.

Findings

LLMs achieve only 28% accuracy on WILT.

Models show varied strengths in hypothesis narrowing and function deduction.

WILT exposes significant reasoning gaps in current LLMs.

Abstract

While large language models have shown impressive capabilities across a wide range of domains, they still encounter significant challenges in reasoning tasks that require gathering evidence over multiple turns and drawing logical conclusions. These challenges present significant obstacles for LLM chat user interfaces, which rely on multi-turn interactions to facilitate effective collaboration. This limitation leads to real-world issues; for example, service chatbots must gather necessary information from customers over multiple turns to diagnose and resolve problems effectively. Despite the multi-turn nature of many real-world LLM use cases, most existing benchmarks rely on carefully curated single-turn tests, which often blur the line between memorization and genuine reasoning. To address this, we introduce the Wason Inductive Logic Test (WILT), a simple yet challenging multi-turn reasoning benchmark designed to resist memorization. WILT is inspired by the Wason 2-4-6 task, where participants must infer a boolean function involving three variables (e.g., $x < y < z$) by proposing test cases (such as $(2, 4, 6)$). In WILT, each test starts from a clean slate, with only the initial instructions provided, preventing models from relying on pre-learned responses. Over several turns, models must interact with the environment by suggesting test cases to narrow the possible hypotheses and ultimately infer the hidden function based on the outcomes. Our findings reveal that LLMs struggle with this task, exhibiting distinct strengths and weaknesses: some are better at narrowing down the hypothesis space by proposing valuable test cases, while others are more adept at deducing the hidden function from observed cases. Despite these variations, the best-performing model achieves only 28% accuracy, highlighting a significant gap in LLM performance on complex multi-turn reasoning tasks.