Hugging Face· Platform

How good are LLMs at fixing their mistakes? A chatbot arena experiment with Keras and TPUs

This blog post describes an experiment testing how well LLMs can fix their mistakes when given feedback in plain English, using a simple calendar API scenario. The author built a chatbot arena using Keras, JAX, and TPUs to interact with multiple LLMs simultaneously.

From the source

I decided to run a little test with today's LLMs. A super-simplified one, to see how effectively LLMs fix their mistakes when you point them out to them.

huggingface.co

Announced

05 Dec 2024 00:00 UTC

Detected

03 Aug 2026 21:12 UTC (+606d after)