when using the llm-rubric assertion the following gets sent to openAI:
[object Object]
Evaluate for output for blah blah
I am using the node.js package for running the eval. I am confirming that llmoutput (even though a JSON) is already a string before I pass it to promptfoo