Gemini 1.5 PRO vs Lllama3-70B-Instruct: TEST

Gemini 1.5 PRO compared to the smaller, new Llama 3 70B Instruct model on causal reasoning and logical deduction.

Based on my personal test suite (copied from Stanford Univ) I tested a multitude of causal reasoning prompt with Gemini 1.5 Pro. Unfortunately Gemini 1.5 PRO solved (almost) all my tests. Therefore I have to design more complex test sets, but for the moment in this video my domain specific and therefore un-representative and also statistically non-representative set of query results …

70b causal gemini gemini 1.5 gemini 1.5 pro llama llama 3 prompt reasoning stanford test tests