---
title: "技能原生大模型：面向长程推理基准与训练的技能熵"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.05139v1"
permalink: "https://feed.astralor.com/items/itm_aefad45b7a940d1b5f2de382"
published: "2026-08-05T17:57:16.000Z"
score: 78
channel: "ai"
tags: ["OpenR1-Math","Qwen3","Skill²-Bench","Skill-Entropy RL","技能熵","长程推理"]
---

# 技能原生大模型：面向长程推理基准与训练的技能熵

## AI 点评

> 这篇论文打动我的地方在于，它没有止步于“模型会犯错”，而是把一种难以名状的认知负担——技能切换——做成了可计算的熵值，并直接用于强化学习奖励。对研究训练算法的人来说，这是很干净的信号设计。做基准评测的朋友也可以参考它的分级方法。

## 速览

本研究针对大模型长程推理中的技能切换能力，提出技能熵（Skill Entropy）以量化跨技能切换难度，并构建含558种技能、9个可验证领域的Skill²-Bench基准。评测发现模型在熵越高任务上准确率显著下降。为此提出Skill-Entropy RL，让模型同时预测答案与所用技能，并引入技能熵奖励。在Qwen3-4B和1.7B上分别从34.4%提至68.4%、14.6%提至40.1%，且可复用至OpenR1-Math等现有数据。

## 关键点

- 提出技能熵（Skill Entropy）量化跨技能切换难度，并构建覆盖558种技能、9个领域的Skill²-Bench基准。
- 评测8个前沿与4个开源模型，发现技能切换差距：任务熵越高，准确率越低。
- 推出Skill-Entropy RL，模型每步同时预测答案与技能，奖励融合步骤正确性与技能熵对齐。
- 在Qwen3-4B与1.7B上分别从34.4%提升到68.4%，从14.6%提升到40.1%，并可迁移至OpenR1-Math。

## 全文

Long-horizon reasoning in recent LLMs demands that the model switch between distinct skills inside a reasoning chain, such as first doing a math derivation, then using the result to plan a schedule. We call such problems cross-skill long-horizon tasks: multi-step tasks whose steps require different reasoning skills and depend on earlier outputs. Existing benchmarks often evaluate individual skills, lacking a principled way to measure how well a model switches between skills. We address this gap from both the evaluation and training sides. We introduce Skill Entropy, a measure of the difficulty of switching from one skill to another. We then propose Skill^2-Bench, a benchmark of cross-skill long-horizon tasks built over 558 skills across 9 verifiable and open-ended domains. Each task is assigned a task-level skill-entropy score and grouped into three difficulty levels. Evaluating 8 frontier and 4 open-source models on Skill^2-Bench reveals a skill-switching gap: accuracy decreases on higher-entropy tasks. We then turn skill entropy from a benchmark scale into a training signal. We propose Skill-Entropy RL, an RL framework where the model predicts not only the answer at each step but also the skill used to produce it. The reward combines step-level correctness with a skill-entropy reward that measures the alignment between the model-predicted skill sequence and the gold skill sequence. On Qwen3-4B-Instruct and Qwen3-1.7B, Skill-Entropy RL improves the Skill^2-Bench score from 34.4% to 68.4% and from 14.6% to 40.1%, respectively, outperforming competitive baselines. The same pipeline can be applied to off-the-shelf training data such as OpenR1-Math, indicating that skill entropy is a reusable training signal. Code available at: https://github.com/Gen-Verse/Skill-Entropy-RL

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_aefad45b7a940d1b5f2de382