A multi-agent hierarchical Bayesian model corrects selection bias in LLM user feedback via topic clustering and reweighting with mild priors on the feedback channel to recover accurate aggregate quality estimates.
Sample selection bias as a specification error
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Correcting Selection Bias in Sparse User Feedback for Large Language Model Quality Estimation: A Multi-Agent Hierarchical Bayesian Approach
A multi-agent hierarchical Bayesian model corrects selection bias in LLM user feedback via topic clustering and reweighting with mild priors on the feedback channel to recover accurate aggregate quality estimates.