Athos Reasoning Verifier (ARV): Scalable Normative Reasoning Assessment for the Post‑Labor AI Era

Fuente: Zenodo
Saved in:
Bibliographic Details
Main Author: Valente, Stefano
Format: Recurso digital
Published: Zenodo 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866902032163536896
author Valente, Stefano
author_facet Valente, Stefano
contents <p>The Athos Reasoning Verifier (ARV) is a scalable framework for assessing human reasoning normalized by age and education, using the Italian Equivalent Score model (Spinnler–Tognoni). ARV performs a 30‑second reasoning verification followed by three lightweight Socratic prompts, enabling population‑level access to structured critical thinking.<br>Validation on 1,247 reasoning samples demonstrates 92.3% accuracy (ICC = 0.87) with no detectable cultural bias. ARV provides the first operational infrastructure for democratic reasoning in the total‑AI era, supporting fair comparison across heterogeneous populations and enabling large‑scale reasoning assessment for alignment, governance, and human–AI interaction.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_18547923
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle Athos Reasoning Verifier (ARV): Scalable Normative Reasoning Assessment for the Post‑Labor AI Era
Valente, Stefano
• Reasoning assessment • Equivalent Score normalization • Socratic verification • Cognitive scaling • Democratic reasoning • AI governance • Post‑labor AI era • Athos architecture • SUPREME‑1 integration • Population‑level reasoning
<p>The Athos Reasoning Verifier (ARV) is a scalable framework for assessing human reasoning normalized by age and education, using the Italian Equivalent Score model (Spinnler–Tognoni). ARV performs a 30‑second reasoning verification followed by three lightweight Socratic prompts, enabling population‑level access to structured critical thinking.<br>Validation on 1,247 reasoning samples demonstrates 92.3% accuracy (ICC = 0.87) with no detectable cultural bias. ARV provides the first operational infrastructure for democratic reasoning in the total‑AI era, supporting fair comparison across heterogeneous populations and enabling large‑scale reasoning assessment for alignment, governance, and human–AI interaction.</p>
title Athos Reasoning Verifier (ARV): Scalable Normative Reasoning Assessment for the Post‑Labor AI Era
topic • Reasoning assessment • Equivalent Score normalization • Socratic verification • Cognitive scaling • Democratic reasoning • AI governance • Post‑labor AI era • Athos architecture • SUPREME‑1 integration • Population‑level reasoning
url https://doi.org/10.5281/zenodo.18547923