STRETCH the Boundaries: A Unified Self-Taught Framework for Progressive LLM Evolution

2026-09-16

AI Summary

为解决大语言模型自我提升时能力停滞问题，提出STRETCH框架，通过动态调整挑战难度与模型能力匹配，促进模型持续进化。

Abstract

Large language models (LLMs) often suffer from capability stagnation in self-improvement training because fixed difficulty levels fail to adapt to their evolving proficiency. To address this issue, we propose STRETCH (Self-Taught Reasoning Evolution via Targeted CHallenge), a unified framework inspired by cognitive scaffolding theory. STRETCH introduces a dynamic Stretch Zone mechanism that continuously aligns question difficulty with the model's solving capability. Within a single parameter space, the model alternates between a Scaffolder that generates adaptive, boundary-pushing challenges and a Learner that that optimizes its solving trajectories through reinforcement learning. This dual-loop co-evolution effectively stabilizes training, mitigates reward hacking and promote progressive reasoning growth. Experiments on both negotiation and operation research benchmarks demonstrate that STRETCH consistently outperforms strong prompting and domain-specific baselines. Further scaffolder configuration analysis shows that dynamic difficulty alignment is critical for sustained capability improvement and synchronized reasoning evolution.

Problem

• large language models
• capability stagnation
• self-improvement training
• fixed difficulty levels

Innovation

• dynamic Stretch Zone
• Scaffolder and Learner co-evolution
• reinforcement learning