4 ms·
Could probably be just as dangerous as helpful. You work through some issues, maybe want to forgive or be forgiven to get closure and the LLM says f*k you, no.
by V__ 3y ago
Could probably be just as dangerous as helpful. You work through some issues, maybe want to forgive or be forgiven to get closure and the LLM says f*k you, no.
- jononor 3y agoThey will RLHF* that away, as best as they can. * Reinforcement Learning from Human Feedback