FOLIO: Natural Language Reasoning with First-Order Logic

TL;DR

FOLIO is a new dataset with first-order logic annotations designed to evaluate and improve the logical reasoning capabilities of large language models in natural language understanding tasks.

Contribution

The paper introduces FOLIO, a human-annotated, logically complex dataset with FOL annotations, and benchmarks LLMs' reasoning abilities on it.

Findings

FOLIO challenges current state-of-the-art LLMs, including GPT-4.

FOLIO enables systematic evaluation of NL reasoning and translation.

Benchmark results highlight gaps in LLMs' logical reasoning skills.

Abstract

Large language models (LLMs) have achieved remarkable performance on a variety of natural language understanding tasks. However, existing benchmarks are inadequate in measuring the complex logical reasoning capabilities of a model. We present FOLIO, a human-annotated, logically complex and diverse dataset for reasoning in natural language (NL), equipped with first-order logic (FOL) annotations. FOLIO consists of 1,430 examples (unique conclusions), each paired with one of 487 sets of premises used to deductively reason for the validity of each conclusion. The logical correctness of the premises and conclusions is ensured by their FOL annotations, which are automatically verified by an FOL inference engine. In addition to the main NL reasoning task, NL-FOL pairs in FOLIO constitute a new NL-FOL translation dataset. Our experiments on FOLIO systematically evaluate the FOL reasoning ability of supervised fine-tuning on medium-sized language models. For both NL reasoning and NL-FOL translation, we benchmark multiple state-of-the-art language models. Our results show that a subset of FOLIO presents a challenge for one of the most capable {Large Language Model (LLM)} publicly available, GPT-4.