MultiZebraLogic: A Multilingual Logical Reasoning Benchmark
This provides a new benchmark for evaluating logical reasoning in LLMs across languages, though it is incremental as it extends existing puzzle-based evaluation methods.
The authors tackled the need for multilingual logical reasoning benchmarks by creating MultiZebraLogic, a dataset of zebra puzzles in nine Germanic languages, finding that puzzle size and red herrings significantly affect model accuracy, with o3-mini's accuracy decreasing by 15±7% when red herrings are included.
Measuring the full abilities of large language models (LLMs) requires benchmarks representing multiple tasks. We aim to create large, high-quality datasets for comparison of logical reasoning skills across several languages and of suitable difficulty for LLMs of various reasoning ability. We explore multiple ways of increasing difficulty. We generate zebra puzzles in multiple languages, themes, sizes and including 14 different clue types and 8 red herring types (uninformative clues). We find puzzle sizes 2x3 and 4x5 are sufficiently challenging for GPT-4o mini (a non-reasoning model) and o3-mini (a reasoning model), respectively. Including 5 red herrings decreases o3-mini puzzle-level accuracy on 4x5 puzzles by 15$\pm$7 %. Scores of o3-mini on 4x5 puzzles are not significantly affected by use of English vs. Danish or the common houses theme vs. the country-specific smoerrebroed theme. We find no correlation between difficulty and the selected clue types. Datasets of 128+1024 puzzles are published as MultiZebraLogic in each of nine Germanic languages for sizes 2x3 and 4x5. We publish code for puzzle generation, designed for adaptablity into more languages and themes.