The Magic of IF: Investigating Causal Reasoning Abilities in Large Language Models of Code

TL;DR

This paper investigates whether large language models trained on code (Code-LLMs) have enhanced causal reasoning abilities, finding that code prompts significantly improve their performance in causal reasoning tasks compared to text-only models.

Contribution

The study demonstrates that Code-LLMs with code prompts outperform text-only models in causal reasoning and highlights the importance of programming structure in prompt design.

Findings

Code-LLMs with code prompts excel in causal reasoning tasks.

Programming structure in prompts is crucial for performance.

Code-LLMs are robust to format perturbations.

Abstract

Causal reasoning, the ability to identify cause-and-effect relationship, is crucial in human thinking. Although large language models (LLMs) succeed in many NLP tasks, it is still challenging for them to conduct complex causal reasoning like abductive reasoning and counterfactual reasoning. Given the fact that programming code may express causal relations more often and explicitly with conditional statements like ``if``, we want to explore whether Code-LLMs acquire better causal reasoning abilities. Our experiments show that compared to text-only LLMs, Code-LLMs with code prompts are significantly better in causal reasoning. We further intervene on the prompts from different aspects, and discover that the programming structure is crucial in code prompt design, while Code-LLMs are robust towards format perturbations.