[ashish1600]

Does this technique only work for certain prompts?

[sourabh03agr]

The authors claim that it works well for a wide variety of cases. They have defined a novel categorisation of these evaluations which help guide LLMs to generate relevant assertions