OPT-R: Exploring the Role of Explanations in Finetuning and Prompting for Reasoning Skills of Large Language Models

TL;DR

This study investigates how explanations influence the reasoning abilities of large language models, specifically OPT, through finetuning and prompting, revealing nuanced effects across different reasoning skills and model configurations.

Contribution

The paper provides a comprehensive analysis of the impact of explanations during finetuning and prompting on LLM reasoning, highlighting when explanations are beneficial or negligible.

Findings

Explanations during finetuning have little impact on performance.

Including explanations in prompting slightly improves accuracy.

Numerical and analogical reasoning benefit most from explanations.

Abstract

In this paper, we conduct a thorough investigation into the reasoning capabilities of Large Language Models (LLMs), focusing specifically on the Open Pretrained Transformers (OPT) models as a representative of such models. Our study entails finetuning three different sizes of OPT on a carefully curated reasoning corpus, resulting in two sets of finetuned models: OPT-R, finetuned without explanations, and OPT-RE, finetuned with explanations. We then evaluate all models on 57 out-of-domain tasks drawn from the SUPER-NATURALINSTRUCTIONS benchmark, covering 26 distinct reasoning skills, utilizing three prompting techniques. Through a comprehensive grid of 27 configurations and 6,156 test evaluations, we investigate the dimensions of finetuning, prompting, and scale to understand the role of explanations on different reasoning skills. Our findings reveal that having explanations in the fewshot exemplar has no significant impact on the model's performance when the model is finetuned, while positively affecting the non-finetuned counterpart. Moreover, we observe a slight yet consistent increase in classification accuracy as we incorporate explanations during prompting and finetuning, respectively. Finally, we offer insights on which skills benefit the most from incorporating explanations during finetuning and prompting, such as Numerical (+20.4%) and Analogical (+13.9%) reasoning, as well as skills that exhibit negligible or negative effects.