Did You Read the Instructions? Rethinking the Effectiveness of Task Definitions in Instruction Learning

TL;DR

This paper investigates how task definitions influence instruction learning in large language models, revealing that minimal, key information suffices and proposing strategies to enhance model understanding and performance.

Contribution

It introduces an automatic method to compress task definitions and proposes two strategies to improve instruction comprehension in LLMs, leading to significant performance gains.

Findings

Model performance drops mainly when removing task output information.

60% of task definition tokens can be removed without performance loss.

Structured key information and meta-tuning improve instruction following by 4.2 Rouge-L.

Abstract

Large language models (LLMs) have shown impressive performance in following natural language instructions to solve unseen tasks. However, it remains unclear whether models truly understand task definitions and whether the human-written definitions are optimal. In this paper, we systematically study the role of task definitions in instruction learning. We first conduct an ablation analysis informed by human annotations to understand which parts of a task definition are most important, and find that model performance only drops substantially when removing contents describing the task output, in particular label information. Next, we propose an automatic algorithm to compress task definitions to a minimal supporting set of tokens, and find that 60\% of tokens can be removed while maintaining or even improving model performance. Based on these results, we propose two strategies to help models better leverage task instructions: (1) providing only key information for tasks in a common structured format, and (2) adding a meta-tuning stage to help the model better understand the definitions. With these two strategies, we achieve a 4.2 Rouge-L improvement over 119 unseen test tasks.