And lest you mistake the author for making just a mild error:how does the LLM know when to stop? How does it know when i...

And lest you mistake the author for making just a mild error:how does the LLM know when to stop? How does it know when it is correct? The astounding answer is that “correctness” and “completeness” and “cohesiveness” are vectors in this space, too. Any correct answer shares the same “correctness” direction with all other factually true statements. He literally thinks that the process of LLM training somehow distills a real-truth value that is a position in latent space.Can you imagine! We'd have a perfect oracle, infinite computational power!This guy gets paid to write books!#LLM #AI #writing #badWriting

Read Original

Related