Submit model to LMSYS' MT Bench
#6
by
lewtun
HF staff
- opened
Hi folks, congrats on training a killer model and releasing the data to go with it 🔥! As an Aussie, I also approve of the name 😁
Now that we have a model with comparable performance to ChatGPT on academic benchmarks, it would be very interesting to submit it to MT Bench so that we can also learn if it is great at multi-turn chat. There's a nice guide here if someone wants to tackle this: https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge#mt-bench