A self-training loop that turns successfully answered queries into evolving capability profiles improves LLM-based task routing by up to 17 percentage points on a proprietary enterprise dataset.
Preprint, arXiv:2410.07490
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
FlyRoute: Self-Evolving Agent Profiling via Data Flywheel for Adaptive Task Routing
A self-training loop that turns successfully answered queries into evolving capability profiles improves LLM-based task routing by up to 17 percentage points on a proprietary enterprise dataset.