Definitely: and just giving the LLM context before correcting (in this case the title and description of the video, often written by a person) creates much better transcripts.