# Distilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict It

**Authors:** Shiping Yang, Chuanqi Tang, Hossein Hajimirsadeghi, Layla El Asri, Mohammad Amin Shabani, Angel X. Chang  
**Published:** 28 Oct 2026  
**Links:** [arXiv](https://arxiv.org/abs/2610.39020) · [PDF](https://arxiv.org/pdf/2610.39020) · [Code](https://github.com/maybenotime/opd-esc) · [This page](https://omen-bit.github.io/ai-paper-daily/paper/2610.39020.html)

## Abstract

On-policy distillation (OPD) is increasingly adopted in the post-training of large language models. However, a stronger teacher may fail to improve the student. Prior work attributes this to a lack of new knowledge, but what that knowledge is remains unclear. We show that OPD learns not only \textit{how to branch}, i.e., which candidate tokens to consider, but also \textit{where to branch}, i.e., where to keep reasoning paths open. The existing token-overlap metric captures only the former. We introduce the entropy shape, the sequence of next-token entropies, to characterize the latter, and propose the Entropy Shape Correlation (ESC) to measure student-teacher agreement in entropy shape. Tracking ESC throughout training and intervening on a teacher's entropy while preserving its token preferences, we empirically show that \textit{where to branch} causally drives OPD gains, independently of \textit{how to branch}. Finally, we formulate OPD outcome prediction as a teacher selection task and establish an evaluation protocol for it, under which pre-OPD ESC selects the best teacher for both students and outperforms existing predictors. Code is available at https://github.com/maybenotime/opd-esc.

---

Notes written by AI from the paper, numbers copied from the abstract. Check the paper before you rely on them. From AI Papers Daily.