---
name: "RecoveringLLMPersona"
description: "Use for Recovering LLM-Persona Accuracies from Unlabeled Votes, a paired GitHub and Zenodo publication with DOI 10.5281/zenodo.20498700."
tags: ["algebraic-evaluation", "ntqr", "unsupervised-evaluation", "evaluation-on-unlabeled-data", "llm-as-judge", "error-independent-evaluation", "ensemble-evaluability", "constant-classifier"]
---

# Recovering LLM-Persona Accuracies from Unlabeled Votes

## Instructions

Use this skill when working with the publication **Recovering LLM-Persona Accuracies from Unlabeled Votes** or its paired release artifacts.

1. Ground citations in DOI `10.5281/zenodo.20498700`.
2. Treat the Zenodo record as the archival source and the GitHub release as the executable/source release.
3. Keep release tag `v1.0.0` and repository `docxology/ntqr_llm` linked when updating catalog surfaces.

## Key Concepts

- **algebraic evaluation**
- **NTQR**
- **unsupervised evaluation**
- **evaluation on unlabeled data**
- **LLM-as-judge**
- **error-independent evaluation**
- **ensemble evaluability**
- **constant classifier**
- **AI safety warning light**
- **reproducible research**
- **answer-key-free recovery**
- **local large language models**

## Prerequisites

- Familiarity with the source repository and Zenodo record.
- Awareness that new versions may update both GitHub and Zenodo surfaces.

## Related

- [README.md](README.md)
- [Full Bibliography](../../pages/BIBLIOGRAPHY.md)