Can you explain how/if this gradient descent concept applies to reasoning or problem solving vs random data generation? I just don’t really understand how this impacts the majority uses of LLMs ...
Age ≥ 18 He/Him Pan(?) Idiot, Asshole cant do anything If I’m talking weirdly/apologising a lot I’m on my meds Go support my friend (@Bluess555) ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results