Shu, Y., Wei, C., Lin, H., Qiu, S., & Xiong, H. (2026). Reference-Sampled Boltzmann Projection for KL-Regularized RLVR: Target-Matched Weighted SFT, Finite One-Shot Gaps, and Policy Mirror Descent.
Style de citation Chicago (17e éd.)Shu, Yao, Chenxing Wei, Hongbin Lin, Shuang Qiu, et Hui Xiong. Reference-Sampled Boltzmann Projection for KL-Regularized RLVR: Target-Matched Weighted SFT, Finite One-Shot Gaps, and Policy Mirror Descent. 2026.
Style de citation MLA (9e éd.)Shu, Yao, et al. Reference-Sampled Boltzmann Projection for KL-Regularized RLVR: Target-Matched Weighted SFT, Finite One-Shot Gaps, and Policy Mirror Descent. 2026.
Attention : ces citations peuvent ne pas être correctes à 100%.