Unpacking the HF in RLHF

Some observations on a few ways different people actually gather feedback from humans in practice to improve LLMs. Sure I've missed some here, so let me know.