Computational · Paper · 2026

Recovering LLM-Persona Accuracies from Unlabeled Votes

Daniel Ari Friedman

Zenodo

Catalog Row152
Citation KeyFriedman2026RecoveringLLMPersonaAccuracies152
Paper FolderAvailable
Platform availability

Overview

Extracted from the local paper documentation when available.

Algebraic (NTQR) evaluation infers how accurate a group of noisy classifiers was on a finite test using only their responses — no answer key. We test this end to end on real large language models. Three trader "personas" (optimistic, neutral, pessimistic), instantiated as system prompts, each make a binary bullish/bearish call on the same 64 market scenarios; we run the identical trio through six...

algebraic evaluationNTQRunsupervised evaluationevaluation on unlabeled dataLLM-as-judgeerror-independent evaluationensemble evaluabilityconstant classifierAI safety warning lightreproducible researchanswer-key-free recoverylocal large language models

Use Notes

Concise findings and methods pulled from README/SKILL documentation.

Findings / Concepts
  • Algebraic (NTQR) evaluation infers how accurate a group of noisy classifiers was on a finite test using only their responses — no answer key.
  • We test this end to end on real large language models.
Methods / Techniques
  • Software pipeline design
  • Data-driven analysis

Citation

Plain-text citation for quick reuse.

Friedman, Daniel Ari. 2026. Recovering LLM-Persona Accuracies from Unlabeled Votes. Zenodo. DOI: 10.5281/zenodo.20498699. URL: https://doi.org/10.5281/zenodo.20498699.

Primary source Documentation Full Text Image Gallery Source repository BibTeX

Related in Computational

Other catalogued works in the same domain.

View all Computational works, software & media →