tracked author

Daixuan Cheng (成岱璇)

First author of LLM-in-Sandbox. Homepage identifies him as a PhD student at GSAI, Renmin University of China, an intern/research student in the GenAI Group at Microsoft Research, and a researcher focused on agentic LLM training.

1 archived notes X: not-found HomepageGitHub

Related Notes

按论文归档时间排序,展示该作者在本站已经出现的材料。

归档

Computer Environments Elicit General Agentic Intelligence in LLMs

LLM in Sandbox 只给模型 shell、文件编辑和完成信号,使部分模型—任务组合获得最高 15.5 个百分点增益,并让 Qwen3 4B 经文件型通用任务强化学习后把平均交互轮次从 23.7 降到 7.0,这些数值来自允许联网的特定环境配置。

待审阅 2601.16206-computer-environments-agentic-intelligence Tool UseAgent RLAgent Workflow