Unnatural Instructions: Tuning Language Models with (Almost) No Human Labor

TL;DR

This paper introduces a large, diverse instruction dataset generated with minimal human effort, enabling effective instruction tuning of language models comparable to manually curated datasets.

Contribution

The authors present Unnatural Instructions, a novel large-scale dataset created with almost no human labor, demonstrating its effectiveness for instruction tuning.

Findings

Training on Unnatural Instructions rivals manually-curated datasets.

Models trained on this dataset outperform T0++ and Tk-Instruct.

Cost-effective data generation method for instruction tuning.

Abstract

Instruction tuning enables pretrained language models to perform new tasks from inference-time natural language descriptions. These approaches rely on vast amounts of human supervision in the form of crowdsourced datasets or user interactions. In this work, we introduce Unnatural Instructions: a large dataset of creative and diverse instructions, collected with virtually no human labor. We collect 64,000 examples by prompting a language model with three seed examples of instructions and eliciting a fourth. This set is then expanded by prompting the model to rephrase each instruction, creating a total of approximately 240,000 examples of instructions, inputs, and outputs. Experiments show that despite containing a fair amount of noise, training on Unnatural Instructions rivals the effectiveness of training on open-source manually-curated datasets, surpassing the performance of models such as T0++ and Tk-Instruct across various benchmarks. These results demonstrate the potential of model-generated data as a cost-effective alternative to crowdsourcing for dataset expansion and diversification.