Ability to manipulate the slope "m" of linear function used to boost relevance scores

Hi there, I have a use case where I need to augment the text based relevance scoring (such as TF-IDF) with a linear function. However, when I went through the Elasticsearch documentation, nowhere does it mention that the slope "m" of the linear function can be tuned according to my needs.

I really need to be able to tune the slope. How can I do this?