@misc{cai2025backtrackingenoughexploringinterplay,
title={How Much Backtracking is Enough? Exploring the Interplay of SFT and RL in Enhancing LLM Reasoning}, 
author={Hongyi James Cai and Junlin Wang and Xiaoyin Chen and Bhuwan Dhingra},
year={2025},
eprint={2505.24273},
archivePrefix={arXiv},
primaryClass={cs.AI},
url={https://arxiv.org/abs/2505.24273}, 
}

@article{Huang2024enhancing,
title  = {Enhancing Large Language Models' Situated Faithfulness to External Contexts},
author = {Yukun Huang and Sanxing Chen and Hongyi Cai and Bhuwan Dhingra},
journal={arXiv preprint arXiv:2410.14675},
year={2024}
}