Prompt/Codebase for reasoning tasks

[varunsingh3000]

Hi, first of all thank you so much for the amazing work and paper.

I was actually interested in evaluating RCI across different reasoning tasks and was wondering about the prompt used in the paper for reasoning. Is it the prompt text mentioned in Figure 2 or was there any additional info in the instructions? The details mentioned in the appendix are not completely clear, for e.g. do you pass all the implicit feedback generated back to the LLM or each individual feedback is passed through a separate call to the LLM without the previous feedback in the prompt? Any chance of releasing the RCI codebase for the reasoning tasks?

Thank you for your time.