metadata

license: mit
task_categories:
  - question-answering
  - text-classification
language:
  - en
arxiv: 2501.14851

JustLogic: A Comprehensive Benchmark for Evaluating Deductive Reasoning in Large Language Models

[Paper] [Github]

JustLogic is a deductive reasoning datataset that is

highly complex, capable of generating a diverse range of linguistic patterns, vocabulary, and argument structures;
prior knowledge independent, eliminating the advantage of models possessing prior knowledge and ensuring that only deductive reasoning is used to answer questions; and
capable of in-depth error analysis on the heterogeneous effects of reasoning depth and argument form on model accuracy.

Dataset Format

premises: List of premises in the question, in the form of a Python list.
paragraph: A paragraph consisting of the above premises. This is given as input to models.
conclusion: The expected conclusion of the given premises.
question: The statement in which models must determine its truth-value.
label: True | False | Uncertain
arg: The argument structure
statements: Matching symbols in arg to their corresponding natural language statements.
depth: The argument depth of the given question

Dataset Construction

JustLogic is a synthetically generated dataset. The script to construct your own dataset can be found in the Github repo.

Citation

@article{chen2025justlogic,
  title={JustLogic: A Comprehensive Benchmark for Evaluating Deductive Reasoning in Large Language Models},
  author={Chen, Michael K and Zhang, Xikun and Tao, Dacheng},
  journal={arXiv preprint arXiv:2501.14851},
  year={2025}
}