I found the issue. The issue wasn't with the dataset format it was with the LLM I used, it wasn't returning the correct output (a value of 0 or 1) , that's why it was giving me RagasOutputParserException. To fix it I tried different models and decreased the number of returned documents from 10 to 5.