S 2 -MAD {\text{S}^{2}\text{-MAD}} : Breaking the Token Barrier to Enhance Multi-Agent Debate Efficiency
Yuting Zeng, Weizhe Huang, Lei Jiang, Tongxuan Liu, Xitai Jin, Chen Tianying Tiana, Jing Li, Xiaohua Xu · 2025
Read S 2 -MAD {\text{S}^{2}\text{-MAD}} : Breaking the Token Barrier to Enhance Multi-Agent Debate Efficiency by Yuting Zeng, Weizhe Huang, Lei Jiang, Tongxuan Liu, Xitai Jin, Chen Tianying Tiana, Jing Li, Xiaohua Xu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-04790.txt.
Description
Yuting Zeng and co-authors cut the cost of debates between large language models, in work from twenty twenty-five. In such debates, several models argue in rounds and reach better answers on hard math and logic tasks. More rounds help, but the word count and cost climb fast.
Their method trims weak or empty exchanges from the debate, so agents share less waste. Tests across data sets and models kept the loss in score under two parts in a hundred. At the same time, word cost fell by up to about ninety-four parts in a hundred. The plan makes model debates cheaper to run at scale. It keeps most of the gain that debate brings.
Catalog record
| CREATOR | Yuting Zeng, Weizhe Huang, Lei Jiang, Tongxuan Liu, Xitai Jin, Chen Tianying Tiana, Jing Li, Xiaohua Xu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 48 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-04790