Saved in:
Bibliographic Details
Main Authors: Cui, Jin, Guo, Jiaqi, Zhou, Jiepeng, Yang, Ruixuan, Lu, Jiayi, Xu, Jiajun, Song, Jiangcheng, Zhao, Boran, Ren, Pengju
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.03717
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915712837091328
author Cui, Jin
Guo, Jiaqi
Zhou, Jiepeng
Yang, Ruixuan
Lu, Jiayi
Xu, Jiajun
Song, Jiangcheng
Zhao, Boran
Ren, Pengju
author_facet Cui, Jin
Guo, Jiaqi
Zhou, Jiepeng
Yang, Ruixuan
Lu, Jiayi
Xu, Jiajun
Song, Jiangcheng
Zhao, Boran
Ren, Pengju
contents While Large Language Models (LLMs) have emerged with remarkable capabilities in complex tasks through Chain-of-Thought reasoning, practical resource constraints have sparked interest in transferring these abilities to smaller models. However, achieving both domain performance and cross-domain generalization remains challenging. Existing approaches typically restrict students to following a single golden rationale and treat different reasoning paths independently. Due to distinct inductive biases and intrinsic preferences, alongside the student's evolving capacity and reasoning preferences during training, a teacher's "optimal" rationale could act as out-of-distribution noise. This misalignment leads to a degeneration of the student's latent reasoning distribution, causing suboptimal performance. To bridge this gap, we propose MIND, a capability-adaptive framework that transitions distillation from passive mimicry to active cognitive construction. We synthesize diverse teacher perspectives through a novel "Teaching Assistant" network. By employing a Feedback-Driven Inertia Calibration mechanism, this network utilizes inertia-filtered training loss to align supervision with the student's current adaptability, effectively enhancing performance while mitigating catastrophic forgetting. Extensive experiments demonstrate that MIND achieves state-of-the-art performance on both in-distribution and out-of-distribution benchmarks, and our sophisticated latent space analysis further confirms the mechanism of reasoning ability internalization.
format Preprint
id arxiv_https___arxiv_org_abs_2601_03717
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MIND: From Passive Mimicry to Active Reasoning through Capability-Aware Multi-Perspective CoT Distillation
Cui, Jin
Guo, Jiaqi
Zhou, Jiepeng
Yang, Ruixuan
Lu, Jiayi
Xu, Jiajun
Song, Jiangcheng
Zhao, Boran
Ren, Pengju
Computation and Language
While Large Language Models (LLMs) have emerged with remarkable capabilities in complex tasks through Chain-of-Thought reasoning, practical resource constraints have sparked interest in transferring these abilities to smaller models. However, achieving both domain performance and cross-domain generalization remains challenging. Existing approaches typically restrict students to following a single golden rationale and treat different reasoning paths independently. Due to distinct inductive biases and intrinsic preferences, alongside the student's evolving capacity and reasoning preferences during training, a teacher's "optimal" rationale could act as out-of-distribution noise. This misalignment leads to a degeneration of the student's latent reasoning distribution, causing suboptimal performance. To bridge this gap, we propose MIND, a capability-adaptive framework that transitions distillation from passive mimicry to active cognitive construction. We synthesize diverse teacher perspectives through a novel "Teaching Assistant" network. By employing a Feedback-Driven Inertia Calibration mechanism, this network utilizes inertia-filtered training loss to align supervision with the student's current adaptability, effectively enhancing performance while mitigating catastrophic forgetting. Extensive experiments demonstrate that MIND achieves state-of-the-art performance on both in-distribution and out-of-distribution benchmarks, and our sophisticated latent space analysis further confirms the mechanism of reasoning ability internalization.
title MIND: From Passive Mimicry to Active Reasoning through Capability-Aware Multi-Perspective CoT Distillation
topic Computation and Language
url https://arxiv.org/abs/2601.03717