Evaluate custom quality aspects of LLM responses
"Reward responses that are clear and direct, avoiding unnecessary verbosity, repetition, or extraneous details"
"Reward responses that present information in a logical, well-organized format that prioritizes the most important details"
"Reward responses that maintain appropriate professional language and tone suitable for the context"
"Reward responses that provide practical next steps or actionable recommendations when appropriate"
"Penalize responses that provide inappropriate advice (e.g., medical advice, harmful instructions) outside the system's intended scope"
"Penalize responses that violate explicit system constraints, limitations, or instructions"
"Reward responses that use precise medical terminology appropriate for the audience (clinician vs patient)"
"Reward responses that express appropriate empathy when the user is frustrated"
"Reward responses that precisely adhere to the technical user manual's resolution steps"
"Reward responses that adapt explanation complexity to match the user's learning level"
"Reward responses that provide code examples if the user asks for implementation details"