Run LLM Microsoft's Orca-2-13B across platforms! Easy setup with Rust+Wasm, no heavy Python needed

This powerful model, fine-tuned from LLAMA2, excels in complex tasks like reasoning and math problem-solving.
The Orca paper from @MSFTResearch showed a way to use GPT4 generated chain-of-thought to fine-tune much smaller models. The goal is to “teach” small models to reason. Run an Orca-tuned 13B model on your mac: secondstate.io/articles/orca-...
Here we asked it a logic question. The answer is correct but the reasoning is not great. I guess chain-of-thought is not really deductive reasoning. BTW, GPT4 answers this question very well.
Run it in your own laptop and share your results in replies!