ChatGPT outputting HTML code

Sample:

The difference between FP16, Q8, and Q4 comes down to a trade-off between model accuracy, memory usage, and inference speed.

It's no longer converting the HTML tags or showing code boxes.