UCLA-AGI / zephyr-7b-sft-full-SPIN-iter1

Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models (https://arxiv.org/abs/2401.01335)