@article{inspounlockingintrinsicselfreflectionfor, title = {InSPO: Unlocking Intrinsic Self-Reflection for LLM Preference Optimization}, author = {Yu Li and Tian Lan and Zhengling Qi}, year = {2025}, eprint = {2512.23126}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2512.23126}, }