Overview
Large Language Models are sensitive to prompt phrasing. Prompt Mining helps discover better templates that occur more frequently in the corpus or are clearer to the model. Here are examples from the paper mapping manual prompts to mined prompts:
We implement a lightweight approach with Instructor to extract clearer prompt templates.
Example
References
- How Can We Know What Language Models Know? (https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00324/96460/How-Can-We-Know-What-Language-Models-Know)