top of page
POCKETP AI INSTITUTE RESEARCH
Does a Structured Moral-Reasoning RAG Improve Rubric-Scored Moral Reasoning in LLM Responses?
Independent Research Report — Not Peer Reviewed
DOI: 10.5281/zenodo.22756832
A Paired, Blinded, Multi-Judge Evaluation
A 300-question blinded evaluation testing whether a structured Moral Reasoning RAG improves measured moral-reasoning performance in the same underlying large language model.
bottom of page