Grader / scorer
Whatever decides how good an output is: code-based, human, or LLM-as-judge.
Whatever decides how good an output is: code-based, human, or LLM-as-judge. (Mod 4)
Whatever decides how good an output is: code-based, human, or LLM-as-judge.
Whatever decides how good an output is: code-based, human, or LLM-as-judge. (Mod 4)