InferenceDynamics: Adaptive LLM Routing through Structured Capability and Knowledge Profiling
TL;DR
This work proposes InferenceDynamics, a flexible and scalable multi-dimensional routing framework by modeling the capability and knowledge of models, and demonstrates its effectiveness and generalizability in group-level routing using modern benchmarks including MMLU-Pro, GPQA, BigGen-Bench, and LiveBench.
Abstract
Large Language Model (LLM) routing is a pivotal technique for navigating a diverse landscape of LLMs, enabling the selection of the best-performing LLMs for specific user queries while balancing performance and cost. However, current routing approaches often face limitations in scalability when dealing with a large pool of specialized LLMs, or in their adapt-ability to extending model scope and evolving capability domains. To overcome those challenges, we propose InferenceDynamics , a flexible and scalable multi-dimensional routing framework by modeling the capability and knowledge of models. We operate it on our comprehensive dataset RouteMix , and demonstrate its effectiveness and generalizability in group-level routing using modern benchmarks including MMLU-Pro, GPQA, BigGen-Bench, and LiveBench, showcasing its ability to identify and leverage top-performing models for given tasks, leading to superior outcomes with cost efficiency. The broader adoption of InferenceDynamics can empower users to harness the full specialized potential of the LLM ecosystem, and our code are publicly available at https://github.com/HKUST-KnowComp/InferenceDynamics.