2 ms·
Sure, I understand that. I remember reading speculation that q* was an attempt to combine LLM training with reinforcement learning. My naive reading is that t
by blurbleblurble 3y ago
Sure, I understand that. I remember reading speculation that q* was an attempt to combine LLM training with reinforcement learning. My naive reading is that this result is somewhat similar and am wondering how it compares to the rumored/speculated q* technique.