Synthesizing and Adapting Error Correction Data for Mobile Large Language Model Applications
This work addresses error correction for mobile LLM users, but it is incremental as it builds on existing data synthesis and adaptation methods.
The paper tackled the problem of improving error correction for large language models (LLMs) in mobile typing applications by synthesizing a high-quality dataset and adapting its distribution to match the mobile domain, resulting in enhanced model performance in both offline evaluation and production A/B testing.
Error correction is an important capability when applying large language models (LLMs) to facilitate user typing on mobile devices. In this paper, we use LLMs to synthesize a high-quality dataset of error correction pairs to evaluate and improve LLMs for mobile applications. We first prompt LLMs with error correction domain knowledge to build a scalable and reliable addition to the existing data synthesis pipeline. We then adapt the synthetic data distribution to match the mobile application domain by reweighting the samples. The reweighting model is learnt by predicting (a handful of) live A/B test metrics when deploying LLMs in production, given the LLM performance on offline evaluation data and scores from a small privacy-preserving on-device language model. Finally, we present best practices for mixing our synthetic data with other data sources to improve model performance on error correction in both offline evaluation and production live A/B testing.