December 3, 2025ComputersOpen Access

EvalCouncil: A Committee-Based LLM Framework for Reliable and Unbiased Automated Grading

View Full Paper

Ask AI

Bookmark

Automated grading with EvalCouncil enhances reliability in assessments, addressing common biases and diagnostic limits.
The framework employs a committee structure with a chief arbitrator to maintain grading consistency and accountability.
Assessment analysis utilizes a graph database to track grading provenance and measure evaluation dispersion through MPAD metrics.
The approach highlights a triage policy that clarifies grading standards for low-dispersion and high-dispersion items for better educational outcomes.

Discussion

No takes yet. Share an insight, caveat, or question.

EvalCouncil improves grading reliability and reduces bias in Machine Learning evaluations, indicating a new approach for academic assessments.

Anghel et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

EvalCouncil improves grading reliability and reduces bias in Machine Learning evaluations, indicating a new approach for academic assessments.