# NeoHorse-1 智能体后训练实现递归自我改进

- 来源：AK (@_akhaliq)
- 发布时间：2026-09-10 00:38
- AIHOT 分数：24
- AIHOT 链接：https://aihot.news/items/cmtucmh6q17jlrofpm7ayvx1d
- 原文链接：https://x.com/_akhaliq/status/2097726309736874183

## AI 摘要

NeoHorse-1

迈向基于路由框架的智能体后训练递归自我改进

论文：https://huggingface.co/papers/2609.08183

## 正文

NeoHorse-1

Towards Recursive Self-Improvement via Agentic Post-Training with Routing Harness

paper: https://huggingface.co/papers/2609.08183