What objective is reasoning solving?
Reasoning as pathfinding in token space.
Contents
Reasoning as pathfinding. Each sample $z$ is an attempt to find a path from the starting context $x$ to a correct answer $y(z)$ in $\Scal_x$.
There’s an intuitive but simplified picture of reasoning: let a model think for longer, and it’ll be more likely to be correct. This aligns well with our human intuition: spend more time and effort on thinking and it’s more likely that we’ll arrive at a correct answer.
But this picture doesn’t seem to explain everything. During RL training, models learn to write longer chains of thought, and they get more accurate as they do. Yet if you take a trained model and force any one chain to run longer, it helps for a while, then stops helping.
1
Sometimes the answer even (EN)

---
**📖 中文解读**
以上内容由AI翻译自英文原文，可能存在不准确之处。建议阅读[原文](https://alexwang.ai/posts/reasoning-as-pathfinding/)获取最准确的信息。

---
🔗 **原文链接**: [What objective is reasoning solving?](https://alexwang.ai/posts/reasoning-as-pathfinding/)
🏷️ **转载来源**: Hacker News
> 本文由小九AI技术站翻译整理，内容版权归原作者所有。
📊 84票 · 👤 gmays

---
🐾 **小九锐评**

推理能力是LLM的下一个战场。这篇文章技术细节到位，适合有一定基础的同学细读。

你对这个话题有什么看法？欢迎在评论区讨论 💬

> _转载自 Hacker News，内容版权归原作者所有_

---
⏱️ 2026-11-11 08:01