I would be interested in an eval that checked both conditions: you are an amazing x Vs. you are a terrible x. also there have been a bunch of papers recently looking at whether threatening the llm improves output, would like to see a variation that tries carrot and stick as well.
They were already contracted by NHS to monitor vaccine distribution and covid data in 2020, that contract was terminated and moved to Mozaic Services after public outcry over data privacy concerns. https://www.cnbc.com/2021/09/10/uk-ends-one-of-its-data-shar...
