RouteNLP: Closed-Loop LLM Routing with Conformal Cascading and Distillation Co-Optimization
Dongxin Guo, Jikun Wu, Siu Ming Yiu · 2026
Read RouteNLP: Closed-Loop LLM Routing with Conformal Cascading and Distillation Co-Optimization by Dongxin Guo, Jikun Wu, Siu Ming Yiu (2026) 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-2604-23577.txt.
Description
Guo, Wu, and Yiu present RouteNLP, a closed-loop router that sends each language-model query to the cheapest tier that can handle it well. A difficulty-aware router, confidence-calibrated cascading, and a loop that distils lessons from escalation failures work together to cut cost. In an eight-week enterprise pilot, inference costs fell by well over half while acceptance stayed high and worst-case latency dropped sharply.
Catalog record
| CREATOR | Dongxin Guo, Jikun Wu, Siu Ming Yiu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45238 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.23577