Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.29167

TeachBack: Learning to Teach via Bidirectional Preference Optimization for On-Policy Self-Distillation

Linfei Ji·Jiayun Li·Haozhao Wang·Siyu Jiang·Zehao Wang·Longtao Xiao·Yu-Jie Yuan·Ruixuan Li
View original source ↗Read source document ↗Export metadata ↓

Abstract

On-policy self-distillation (OPSD) enables a language model to learn from a Teacher instantiated from the same model but given access to privileged information, such as a reference solution. We find that access to sufficient privileged information allows the Teacher to generate highly accurate trajectories; however, accuracy alone does not guarantee that these trajectories are suitable for Student learning. Direct access to a complete solution may encourage shortened, reference-conditioned derivations that omit the exploration, verification, and self-correction required to solve problems without privileged information. We introduce TeachBack, a bidirectional preference-learning framework for on-policy self-distillation that learns not only from the Teacher but also how the Teacher should teach. TeachBack shapes the Teacher's generation behavior by controlling the length of the reference-reasoning suffix injected into the prompt and by varying the prompt style. While maintaining comparable Teacher accuracy, it encourages the Teacher to explore and re-solve problems, thereby reducing its reliance on the reference solution. At each rollout refresh, the shared policy generates both a privileged Teacher trajectory and a problem-only Student trajectory. A pairwise preference objective uses privileged Teacher trajectories to improve Student learning, while the reversed preference uses the Student's behavior to shape future Teacher supervision under the privileged condition. Experiments on mathematical and logical reasoning tasks show that TeachBack consistently improves Student performance over existing OPSD methods and related baselines.

Explore related subjects

cs.AIcs.CLcs.LG

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Linfei Ji, Jiayun Li, Haozhao Wang, Siyu Jiang, Zehao Wang, Longtao Xiao, Yu-Jie Yuan, Ruixuan Li. 2026-10-21. TeachBack: Learning to Teach via Bidirectional Preference Optimization for On-Policy Self-Distillation. https://arxiv.org/abs/2610.29167

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

RoSeViT: Role-Separated Vision Transformers for ARC Visual Reasoning

ARC tasks can be formulated as visual transformation problems, where a model must infer a high-level transformation rule from a small set of demonstrations and apply it consistently across a grid. While dense Vision Transformers (ViTs) perform well in this setting, they entangle local visual state and global transformation logic within a shared token representation, which can hinder consistent reasoning. In this paper, we introduce RoSeViT, a role-separated Vision Transformer that explicitly decouples high-level instruction inference from visual workspace computation. RoSeViT uses controller tokens to aggregate transformation-level information and workspace tokens to represent the visual grid, with a structured attention pass in which nonlocal workspace interactions are routed through the controller. This design is motivated by our formal analysis showing that enforcing a shared instruction representation can reduce the effective complexity of visual transformation modeling. RoSeViT outperforms dense ViT baselines under the standard VARC training and evaluation protocol for ARC. In a controlled backbone replacement setting, it improves ARC-1 accuracy from 54.5 to 56.6, and with recurrent refinement reaches 58.8. When integrated into the full system, RoSeViT-Ensemble improves performance from 60.4 to 62.6 on ARC-1. These results demonstrate that explicitly separating instruction and workspace representations leads to more effective visual reasoning on ARC tasks. GitHub repository: https://github.com/lz7fd/RoSeViT.

cs.AI↗
arXiv2026

Learning to Model Optimization Problems from Solver Outcomes

Automating optimization modeling with LLMs is a promising path toward scalable decision intelligence, but existing approaches either rely on agentic pipelines built on closed-source LLMs with high inference latency or fine-tune smaller LLMs using costly process supervision that often overfits to a single solver API. We study how reinforcement learning from inexpensive outcome labels can improve optimization modeling and support reuse across heterogeneous solver environments. To this end, we propose Optimization Modeling from Solver Outcomes (OMSO), which uses reinforcement learning (RL) with a mathematical programming solver as a deterministic, interactive environment. Given a natural-language problem and a target solver, OMSO generates solver-specific code, executes it in a sandboxed harness, and converts execution outcomes into scalar rewards in a closed-loop generate-execute-feedback-update process. This RL formulation learns optimization modeling without process-level supervision and supports cross-solver generalization by switching the solver environment rather than reconstructing solver-specific datasets. Experiments on NL4OPT, MAMO, IndustryOR, and OptiBench across Gurobi, OR-Tools, and COPT show that OMSO matches or outperforms process-supervised SFT, supports zero-shot training-free transfer, direct RL training, and SFT-cold-started RL, and improves modeling reasoning and semantic correctness beyond solver API usage.

cs.AI↗
arXiv2026

Forget Attention: Importance-Aware Attention Is All You Need

Attention retrieves content precisely, but its scores depend only on query--key products, leaving order and recency to positional encodings. Conversely, state-space models (SSMs) such as Mamba-3 track order and recency through input-dependent state transitions that decay and rotate, yet compressing context into a fixed-size state hampers precise retrieval. Hybrids like Samba and Hymba exploit this complementarity only across layers or heads, so SSM transitions never shape attention scores directly. The closest prior work, the Forgetting Transformer (FoX), adds an input-dependent forget gate to the scores but reweights attention by recency alone. We propose SSM-Informed Softmax Attention (SISA), which builds Mamba-3-style state transitions, rotation included, directly into the scores alongside FoX's gate. Each logit receives the weight such an SSM would give that token pair: a decayed match between rotated SSM projections. This rotation already encodes relative position, so SISA drops RoPE; restoring it raises loss. SISA outperforms parameter-matched Transformer and FoX in language modeling and length extrapolation. Among the attention models we compare, hybrids included, only SISA benefits from context beyond the training window. Building SSM transitions into the scores yields stronger attention, and our kernels make it faster and leaner than baselines on their official or standard kernels.

cs.AI↗