Improving reasoning models by learning from nearby expert suggestions

Better Supervision Is Nearby: Neighborhood On-Policy Self-Distillation

Computation and LanguageMachine Learning

Summary

Training AI models that solve math problems can be helped by learning not just from one expert solution, but from a group of slightly different expert views nearby. The authors show that these nearby experts provide useful corrections at different steps of solving a problem, giving better guidance to the model. Their method, called Neighborhood On-Policy Self-Distillation, selects the best experts dynamically and improves model accuracy on math benchmarks compared to previous approaches. This approach learns from the full advice of the selected expert, not just the top suggestion, helping models generalize better.

What this means in practice

  • For ai model trainers: Enhance training of mathematical reasoning AI models by incorporating dynamically selected nearby expert corrections for improved accuracy.
  • For automated tutoring developers: Improve AI systems that generate step-by-step math solutions by using richer supervision signals from multiple expert viewpoints during training.

Authors

Xincheng Wei, Yifan Ding, Yoshua Li, Yuquan Lu, Ziheng Li, Yi Lu, Dongsheng Ma, Rongxiang Weng, Xunliang Cai

Abstract

On-policy self-distillation (OPSD) trains mathematical reasoning models using a privileged teacher that sees a reference solution and supervises student-sampled prefixes. Standard OPSD uses one fixed parameter setting at every state, but nearby settings may offer additional supervision. We find that local parameter perturbations reveal complementary reference-aligned corrections under the same reference context. Different experts supply these corrections at different reference positions. Their pool covers more such positions than the unperturbed privileged teacher. We introduce Neighborhood OPSD (N-OPSD) to turn these corrections into supervision at student-visited states. Offline, greedy selection builds a compact pool of frozen experts by rewarding filtered reference-token gains beyond the pool's current best at each position. The highest-peak expert need not provide the best training target. Online routing therefore separates the anchor direction from its level of support. MaxPeak selects the anchor token, and quantile selection chooses among experts whose top token matches it. The student learns from the chosen expert's full next-token distribution through the clipped forward-KL objective inherited from OPSD. We evaluate on AIME 2024, AIME 2025, and HMMT February 2025. Across three independent runs per method, Neighborhood OPSD improves the three-benchmark Average@12 over OPSD by 2.75, 1.67, and 1.94 points on Qwen3-1.7B, 4B, and 8B, respectively. Student-prefix continuations support using the pool beyond the reference trajectories used for selection. Matched ablations support filtered reference-token gains as a selection criterion. Accounting for overlap within the pool and routing by state further improve student accuracy. Inference uses only the distilled student.