When using a predefined LLM-as-a-judge LangChainStringEvaluator of the type score_string and viewing the run of the experiment in the LangSmith console, the score_string evaluator columns display the evaluator name with a “Score_string: ” prefix, then the entire content of the llm-as-a-judge prompt provided for scoring (which is quite long and ugly).
This is unlike the other evaluators which display friendly evaluator names (e.g. output_format_evaluator).
Is there a way to set these score_string LangChainStringEvaluator evaluators user friendly names like this? Or some other workaround to this issue?