LLM-Blender：利用成對排名和生成融合集成大型語言模型

LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusion

June 5, 2023

作者: Dongfu Jiang, Xiang Ren, Bill Yuchen Lin

cs.AI

摘要

我們提出了LLM-Blender，一個整合框架，旨在通過利用多個開源大型語言模型（LLMs）的多樣優勢，實現始終優異的性能。我們的框架包括兩個模塊：PairRanker和GenFuser，解決了對不同示例最優LLMs可能存在顯著差異的觀察。PairRanker採用專門的成對比較方法來區分候選輸出之間的細微差異。它聯合編碼輸入文本和一對候選輸出，使用交叉注意力編碼器來確定優越者。我們的結果表明，PairRanker與基於ChatGPT的排名具有最高的相關性。接著，GenFuser旨在合併排名靠前的候選輸出，通過利用它們的優勢並減輕它們的弱點，生成一個改進的輸出。為了促進大規模評估，我們引入了一個基準數據集MixInstruct，這是多個指令數據集的混合，具有oracle成對比較。我們的LLM-Blender在各種指標上顯著優於單個LLMs和基準方法，確立了實質性的性能差距。

English

We present LLM-Blender, an ensembling framework designed to attain consistently superior performance by leveraging the diverse strengths of multiple open-source large language models (LLMs). Our framework consists of two modules: PairRanker and GenFuser, addressing the observation that optimal LLMs for different examples can significantly vary. PairRanker employs a specialized pairwise comparison method to distinguish subtle differences between candidate outputs. It jointly encodes the input text and a pair of candidates, using cross-attention encoders to determine the superior one. Our results demonstrate that PairRanker exhibits the highest correlation with ChatGPT-based ranking. Then, GenFuser aims to merge the top-ranked candidates, generating an improved output by capitalizing on their strengths and mitigating their weaknesses. To facilitate large-scale evaluation, we introduce a benchmark dataset, MixInstruct, which is a mixture of multiple instruction datasets featuring oracle pairwise comparisons. Our LLM-Blender significantly outperform individual LLMs and baseline methods across various metrics, establishing a substantial performance gap.