ComputeGPT: A computational chat model for numerical problems
This addresses the issue of numerical inaccuracy in language models for users needing reliable computational answers, though it is incremental as it builds on existing chat models with added code execution.
The paper tackles the problem of language models being inaccurate in numerical problems by introducing ComputeGPT, a chat model that converts questions to code, runs it, and returns computed answers, achieving state-of-the-art efficiency on numerical problems.
Language models are not accurate in numerical problems. Their architecture does not allow for anything less than a probabilistic next word. This paper introduces ComputeGPT: an approach of creating a chat model able to answer computational problems through running on-demand code. ComputeGPT converts each question to relevant code, runs the code, and returns the computed answer as part of the chat. We combine this approach with a local browser-based Python interpretation and fine-tuned prompts in order to achieve state-of-the-art efficiency on numerical problems and provide a suitable front-end and safe environment for the code to be executed in.