Contributor
Scale AITextMappingRLHF
I propose a difficult single-choice legal problem with 4 choices and then the model will generate a response with a detailed reasoning. I need to make the model fail at least one step of the reasoning. My task is to correct the wrong logical reasoning and write a justification.
I propose a difficult single-choice legal problem with 4 choices and then the model will generate a response with a detailed reasoning. I need to make the model fail at least one step of the reasoning. My task is to correct the wrong logical reasoning and write a justification.
2024