Dimension evaluated
Specific issues identified
Whether the LLM response was successfully parsed
LLM's reasoning
Score from the LLM (0-1)
Severity from the LLM
Dimension evaluated