AI and LLMs · Junior · Question 43 of 75
What is RLHF?
- 1A rule that limits how fast models can generate text
- 2Tuning a model with human preference feedback
- 3A file format for sharing prompts between teams
Show the answer
Correct answer: Tuning a model with human preference feedback
Explanation
Humans compare answers, a reward model learns their preferences, and the LLM is tuned toward preferred behaviour.