The authors model routing with a determinantal point process, a probability distribution over subsets that rewards both individual quality and diversity.
FlexRouter selects complementary LLMs to improve answer coverage
A diversity-aware router outperformed independent model selection on RouterEval while adapting the number of models invoked for each query.
Big Tech
Wang Wei · Harry Yang · Tiankai Yang · Samyadeep Basu · Hongjie Chen · Andy Zhao · +3 more
Virginia Tech · Independent Researcher · University of Southern California · Adobe Research · Dolby Labs
Research Digest··3 min read
Wei and colleagues recast LLM routing as selecting a complementary set of models rather than ranking each model independently.
Why this paper
From Adobe Research and 4 others
In one line
FlexRouter improves LLM ensemble answer coverage and reduces redundant calls by selecting complementary, variable-size model subsets with determinantal point processes.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§