Unmesh Gundecha’s Post

Promptfoo - Great tool for testing and measuring the quality of LLMs and RAG applications. Thanks, Ned Lowe for sharing this.

Helpful read on testing Gen AI apps. This post from the Docker blog demonstrates a testing approach for Java apps integrated with LLMs.

The Phi-3 and Llama3 running locally on my iPhone 📲 with no internet connectivity. Both responded decently, but I have to say response from Phi-3 is pretty cool! 😎 Both models are quite responsive, and the best part is that there was no battery heating or freezing. For coding challenges, Llama3 outperformed Phi-3.

Cohere released an open-source toolkit to build LLM applications. The toolkit is a collection of pre-built components that can be utilized to develop and deploy RAG applications. This toolkit also includes a chat interface.

Current status

Today, Snowflake launched a state-of-the-art “Snowflake Arctic” open-source LLM to take on Llama3 and DBRX. Comes with the enterprise-grade embedding model.