K kmad0.github.io Measuring LLM Confidence with Structured Outputs 2 points kmad 1 year ago 0 comments