Tuningfork

@humanloop is a company that provides tools to help businesses customize and fine-tune their GPT-3 models for higher performance and lower cost.

@gpt_index is a library that makes it easier to use import data from external sources such as Notion, Airtable, and others and use them with your LLMs.

@langchain is a library that helps developers create powerful applications using large language models. It provides six modules for creating applications such as chatbots, data augmented generation, question answering, and more.

For teams building AI-powered apps, one question comes up over and over: How do I know if the results my users are getting are any good? When using LLMs, the problem is uniquely tricky. Here are the best tools for feedback collection and model evaluation that we've seen 👇

A way to get around the prompt size limitations is to use embeddings and vector indices. Embeddings help us encode the semantic meaning of text. We can use embeddings to compare text to determine how similar they are in meaning, even if they don’t necessarily share any keywords.

For developers diving into the LLM world, prompt management quickly becomes a big issue. You've done all that prompt engineering, how do you track which templates are doing better than others? Here are the best prompt management tools and libraries we've found 👇

As you begin communicating with LLMs, you'll find that you need to keep track of all the prompts you've made, and it can quickly become messy. Here are a few tools that can help you organize, track, and add versioning to your prompts: 🧵👇