CentaurTA Studio reaches up to 92.12% accuracy in open coding and theme construction across three domains by using a two-stage human feedback loop, persistent prompt optimization, and rubric-based early stopping, outperforming baselines with substantial human-LLM agreement.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.HC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
CentaurTA Studio: A Self-Improving Human-Agent Collaboration System for Thematic Analysis
CentaurTA Studio reaches up to 92.12% accuracy in open coding and theme construction across three domains by using a two-stage human feedback loop, persistent prompt optimization, and rubric-based early stopping, outperforming baselines with substantial human-LLM agreement.