TruthLLM

• 

  GOAT Basketball Player Evaluation

Inspiration

We were inspired by the speaker from yesterday, Craig Martell, about how more and more people are blindly trusting AI responses. We are inspired to encourage critical thinking and thoughtful decision-making for LLM users.

What it does

The project is a Google Chrome extension that analyzes the output of a given LLM, and creates a "truthfulness" percentage based on the language it provides in the output.

How we built it

We first created a top-down design of all the features we wanted to add in this extension. Then we created a python file to weight the evaluation metrics, and used PearAI to generate the framework for the Chrome extension.

Challenges we ran into

The biggest challenge was learning to host the code on Google Cloud. Me and my team have very little computer science knowledge, so it was very difficult to build and export the project with the correct APIs.

Accomplishments that we're proud of

This is our first hackathon and the first product we have ever built!

What we learned

We are proud that we were able to learn more about cloud systems and general computer network information.

What's next for TruthLLM

We want to implement the Truth feature for all LLMs, currently it is only available to ChatGPT. It would also be interesting to try and sell the product.

Built With

• google-cloud
• javascript
• json
• python