AgentRec: Agent Recommendation Using Sentence Embeddings Aligned to Human Feedback

TL;DR

AgentRec is a novel, efficient system that recommends the most suitable LLM agent for a task by leveraging sentence embeddings aligned with human feedback, achieving high accuracy and interpretability.

Contribution

This work introduces a new architecture extending SBERT for agent recommendation, incorporating reinforcement learning to align embeddings with human values.

Findings

92.2% top-1 accuracy on test data

Classification time less than 300 milliseconds

Open-sourced synthetic dataset and code

Abstract

Multi-agent systems must decide which agent is the most appropriate for a given task. We propose a novel architecture for recommending which LLM agent out of many should perform a task given a natural language prompt by extending the Sentence-BERT (SBERT) encoder model. On test data, we are able to achieve a top-1 accuracy of 92.2% with each classification taking less than 300 milliseconds. In contrast to traditional classification methods, our architecture is computationally cheap, adaptive to new classes, interpretable, and controllable with arbitrary metrics through reinforcement learning. By encoding natural language prompts into sentence embeddings, our model captures the semantic content relevant to recommending an agent. The distance between sentence embeddings that belong to the same agent is then minimized through fine-tuning and aligned to human values through reinforcement learning from human feedback. This allows the classification of natural language prompts based on their nearest neighbors by measuring the cosine similarity between embeddings. This work is made possible through the generation of a synthetic dataset for agent recommendation, which we have open-sourced to the public along with the code for AgentRec recommendation system at https://github.com/joshprk/agentrec.