Harnessing the Power of Multiple Minds: Lessons Learned from LLM Routing

KV Aditya Srivatsa, Kaushal Kumar Maurya, Ekaterina Kochmar
5/1/2024
cs.CL

No official code found. Best implementation: kvadityasrivatsa/llm-routing (7 stars, MIT, last commit 2024-07-06).

Abstract

With the rapid development of LLMs, it is natural to ask how to harness their capabilities efficiently. In this paper, we explore whether it is feasible to direct each input query to a single most suitable LLM. To this end, we propose LLM routing for challenging reasoning tasks. Our extensive experiments suggest that such routing shows promise but is not feasible in all scenarios, so more robust approaches should be investigated to fill this gap.

View on arXivView PDF

Code Implementations(1)

kvadityasrivatsa/llm-routingCommunity100%
MIT

This repo contains the data, code, and analysis pertaining to our paper -- Harnessing the Power of Multiple Minds: Lessons Learned from LLM Routing.

71Shell, PythonApr 24, 20242 years agoMIT

Similar Papers

Dynamic Model Routing and Cascading for Efficient LLM Inference: A Survey

85% match2026

Routesplain: Towards Faithful and Intervenable Routing for Software-related Tasks

84% match2025

RADAR: Reasoning-Ability and Difficulty-Aware Routing for Reasoning LLMs

84% match2025

BEST-Route: Adaptive LLM Routing with Test-Time Optimal Compute

83% match1 repo662025

DOTS: Learning to Reason Dynamically in LLMs via Optimal Reasoning Trajectories Search

83% match1 repo42024

LLMRank: Understanding LLM Strengths for Model Routing

83% match2025

Cite this paper

@article{srivatsa2024harnessing,
  title  = {Harnessing the Power of Multiple Minds: Lessons Learned from LLM Routing},
  author = {KV Aditya Srivatsa and Kaushal Kumar Maurya and Ekaterina Kochmar},
  year   = {2024},
  eprint = {2405.00467},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2405.00467v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2405.00467.svg)](https://codeofpaper.com/paper/2405.00467)

Discussion