Enabling Ensemble Learning for Heterogeneous Large Language Models with Deep Parallel Collaboration

arXiv:2404.12715v1 Announce Type: new
Abstract: Large language models (LLMs) have shown complementary strengths in various tasks and instances, motivating the research of ensembling LLMs to push the frontier leveraging the wisdom of the crowd. Existing work achieves this objective via training the extra reward model or fusion model to select or fuse all candidate answers. However, these methods pose a great challenge to the generalizability of the trained models. Besides, existing methods use the textual responses as communication media, ignoring …

abstract arxiv collaboration cs.cl enabling ensemble extra fusion instances language language models large language large language models llms research reward model tasks training type via work