Reasoning Level

View as Markdown

A control for how much deliberate chain-of-thought style computation a model should apply.