Scoring, Reasoning, and Selecting the Best! Ensembling Large Language Models via a Peer-Review Process

TL;DR

This paper introduces LLM-PeerReview, an unsupervised ensemble method that selects the best response from multiple large language models using a peer-review-inspired, transparent framework, improving performance across various tasks.

Contribution

It presents a novel, fully unsupervised peer-review-inspired ensemble approach that effectively combines multiple LLM responses with interpretability and adaptability.

Findings

Outperforms the advanced model Smoothie-Global by 6.9% and 7.3% points.

Works effectively across factual recall, math reasoning, and instruction following tasks.

Uses LLM-as-a-Judge and a graphical model-based inference for response selection.

Abstract

We propose LLM-PeerReview, an unsupervised LLM Ensemble method that selects the most ideal response from multiple LLM-generated candidates for each query, harnessing the collective wisdom of multiple models with diverse strengths. LLM-PeerReview is built on a novel, peer-review-inspired framework that offers a transparent and interpretable mechanism, while remaining fully unsupervised for flexible adaptability and generalization. Specifically, it operates in three stages: For scoring, we use the emerging LLM-as-a-Judge technique to evaluate each response by reusing multiple LLMs at hand; For reasoning, we can apply a straightforward averaging strategy or a principled graphical model-based truth inference algorithm to aggregate multiple scores to produce a final score for each response; Finally, the highest-scoring response is selected as the best ensemble output. LLM-PeerReview is conceptually simple and empirically powerful. Our results across four datasets show that the two variants of the proposed approach outperform the advanced model Smoothie-Global by 6.9% and 7.3% points, cross diverse task types including factual recall QA, math reasoning, and instruction following.