新闻
Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B
arXiv cs.AI · 发布于 · 阅读约1分钟
30秒读懂
- 发生了什么
- A controlled experiment shows simple repeated sampling outperforms or matches self-refining methods like Self-Refine and Reflexion when token budgets are equal across 1.5B to 7B models.
- 为何重要
- Matters for engineers choosing inference strategies for math problems and evaluating whether complex reasoning methods justify their computational cost.
- 注意
- Study limited to two math benchmarks with 150 questions each; results may not generalize to other domains, languages, or task types beyond mathematics.
收听本摘要
- language model
- token
The Agent Architect
每周一个模式、一个权衡、一个生产事故案例。为构建智能体系统的人准备的每周简报。
每周一封邮件,一键退订。您的地址仅用于发送简报。