In question answering requiring common sense, language models (e.g., GPT-3) have been used to generate text expressing background knowledge that helps improve performance. Yet the cost of working with such models is very high; in this work, we finetune smaller language models to generate useful intermediate context, referred to here as elaborations. Our framework alternates between updating two language models -- an elaboration generator and an answer predictor -- allowing each to influence the other. Using less than 0.5% of the parameters of GPT-3, our model outperforms alternatives with similar sizes and closes the gap on GPT-3 on four commonsense question answering benchmarks. Human evaluations show that the quality of the generated elaborations is high.
@article{arxiv.2209.01232,
title = {Elaboration-Generating Commonsense Question Answering at Scale},
author = {Wenya Wang and Vivek Srikumar and Hanna Hajishirzi and Noah A. Smith},
journal= {arXiv preprint arXiv:2209.01232},
year = {2023}
}