December 11, 2023:
Marcoroni-7B-v3 has placed
#5 overall and
#1 for 7 billion parameter models on the
Hugging Face Leaderboard!
This is a DPO fine tuned model of
Q-bert/MetaMath-Cybertron-Starling.
We fine-tuned using 32k data generated by GPT-4 and other models.
### Instruction:
<prompt> (without the <>)
### Response:
