Privileged on-policy distillation improves multimodal reasoning by allowing a teacher to evaluate student trajectories using rich, training-only visual evidence. Both models score these trajectories w...
OPD‑Aha is a privileged on‑policy distillation method that improves multimodal reasoning by reconstructing the distillation target from the teacher’s isolated visual preference instead of relying on fragile teacher‑student discrepancies. It suppresses continuations that contradict the image, encouraging students to interrupt flawed reasoning with reflection tokens such as "wait" and "actually." This approach leads to consistent improvements across fine‑grained perception and complex multimodal reasoning benchmarks.
By Chenhao Qiu, Dawei Li, Yechao Zhang, Lei Gong, Zhen Tan
arXiv:2609.36838v1 Announce Type: cross
Abstract: Visual agents solve problems by interleaving reasoning with image operations, and on-policy distillation (OPD) provides guidance from a strong teache...
By Shaohang Wei, Feifan Song, Guangyue Peng, Wenhao Yu, Wei Li, Wen Luo, Yang Xu, Yufan Shen, Luke Mao, Yang Du, Asher Qin, Houfeng Wang
Contrastive On-Policy Distillation (COPD) is a framework that improves on-policy distillation by using a frozen teacher to evaluate student states under two contrasting prompts—one encouraging low reasoning effort and one encouraging high effort. The difference in log‑probabilities between these prompts provides a token‑level advantage signal that guides the student toward more concise and efficient reasoning strategies. Experiments on nine multimodal benchmarks show that COPD reduces reasoning length while maintaining task performance, and the contrastive approach can also be applied to on‑policy self‑distillation, allowing a model to compress its own reasoning without an external teacher.
By Jiacheng Ruan, Jun Tang, Wenzhen Yuan, Ting Liu, Shuai Bai, Dayiheng Liu, Zhibo Yang, Yuzhuo Fu
arXiv:2607. 21556v1 Announce Type: cross Abstract: On-policy self-distillation (OPSD) is promising as it removes the external teacher required by on-policy distillation (OPD), yet it still needs asymmetric information between teacher and student to ensure that the self-teacher provides a stronger learning signal than the student.
By Yijun Liang, Yunjie Tian, Yijiang Li, Yuqi Jia, Furong Huang, Tianyi Zhou, Di Fu
The paper introduces Verifier-Gated Multi-Expert On-Policy Distillation (VG‑OPD), a method that assigns teacher supervision at the token level based on each expert’s counterfactual gain on a specific answer criterion. VG‑OPD localizes supervision where experts disagree most with the student and weights it by criterion importance, integrating this into a gated KL advantage for reinforcement learning. Applied to scientific reasoning, VG‑OPD achieves top performance on seven benchmarks for 4B and 8B models, outperforming prior multi‑teacher distillation approaches.
By Xun Xu, Zaixi Zhang