> Completely different results as LLM is non-deterministic.
You'd need to produce this like 20 times by each model and then do 2x20x20 cross comparisons by both models and ultimately distill the 2x20x20 comparison results into two reports of how they differ.
In this non deterministic computing future, everything else is voodoo, feelings and "vibes".
You'd need to produce this like 20 times by each model and then do 2x20x20 cross comparisons by both models and ultimately distill the 2x20x20 comparison results into two reports of how they differ.
In this non deterministic computing future, everything else is voodoo, feelings and "vibes".