What are Autoraters ?

Large language models (LLMs) produce certain output based on the input. How do we determine if the the output is correct or of good quality ? One way to do this is by employing humans who actually read the output and rate it. This of course is not sc...