Post-Train Your Local AI: Which Method Actually Works

Post-training your LLM: SFT, RL, DPO, OPD, OPSD, and more Distillation methods (via SFT and RLVR) explained. All rights w/ authors: Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models Siyan Zhao† 1 Zhihui Xie 2 Mengchen Liu 3 Jing Huang 3 Guan Pang 3 Feiyu Chen∗,‡ 3 Aditya Grover 1 UCLA 2 HKU 3 Meta Superintelligence Labs. https://github.com/siyan-zhao/OPSD arXiv:2601.18734v3 Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe Yaxuan Li∗1,2, Yuxin Zuo∗†1, Bingxiang He∗†1, Jinqian Zhang1, Chaojun Xiao‡1, Cheng Qian3,