Re-Tuning: Overcoming the Compositionality Limits of Large Language Models with Recursive Tuning

TL;DR

Re-Tuning is a recursive tuning method that enhances large language models' ability to solve compositional tasks by breaking down problems into subproblems, leading to higher accuracy and efficiency.

Contribution

The paper introduces Re-Tuning, a novel recursive tuning approach that improves large language models' performance on compositional tasks by enabling problem decomposition.

Findings

Significantly improves accuracy on compositional tasks

More GPU memory efficient than state-of-the-art methods

Effective on tasks like integer addition, dynamic programming, and parity

Abstract

We present a new method for large language models to solve compositional tasks. Although they have shown strong performance on traditional language understanding tasks, large language models struggle to solve compositional tasks, where the solution depends on solving smaller instances of the same problem. We propose a natural approach to solve compositional tasks recursively. Our method, Re-Tuning, tunes models to break down a problem into subproblems, solve those subproblems, and combine the results. We show that our method significantly improves model performance on three representative compositional tasks: integer addition, dynamic programming, and parity. Compared to state-of-the-art methods that keep intermediate steps towards solving the problems, Re-Tuning achieves significantly higher accuracy and is more GPU memory efficient.