CharED: Character-wise Ensemble Decoding for Large Language Models

Large language models (LLMs) have shown remarkable potential for problem solving, with open source models achieving increasingly impressive performance on benchmarks measuring areas from logical reasoning to mathematical ability. Ensembling models can further improve capabilities across a variety of...

Full description

Saved in:
Bibliographic Details
Published inarXiv.org
Main Authors Gu, Kevin, Tuecke, Eva, Katz, Dmitriy, Horesh, Raya, Alvarez-Melis, David, Yurochkin, Mikhail
Format Paper
LanguageEnglish
Published Ithaca Cornell University Library, arXiv.org 25.06.2024
Subjects
Algorithms
Benchmarks
Cognition & reasoning
Decoding
Inference
Large language models
Online AccessGet full text

Cover

More Information
Summary:Large language models (LLMs) have shown remarkable potential for problem solving, with open source models achieving increasingly impressive performance on benchmarks measuring areas from logical reasoning to mathematical ability. Ensembling models can further improve capabilities across a variety of domains. However, conventional methods of combining models at inference time such as shallow fusion necessitate a shared vocabulary and tokenization, and alternatives like fine-tuning for domain-specific performance are both time consuming and computationally expensive. We therefore present an inference-time ensembling algorithm aimed at "averaging" outputs from multiple LLMs and illustrate its improved performance across multiple domains compared to its constituent models alone. Character-wise ensemble decoding, CharED, finds the marginal distribution of each character for an individual model and performs a weighted average to generate an output, character by character. In coding, math, and toxicity benchmarks, we find our proposed model able to combine complimentary strengths of multiple LLMs, regardless of vocabulary, tokenization, or model size.
ISSN:2331-8422