通过集成学习提升大模型在文本与代码生成中的多样性与质量
Ensemble Learning for Large Language Models in Text and Code Generation: A Survey
- 将多个大模型按权重融合、知识融合等方式集成，增强输出多样性
- 集成后模型在文本和代码生成任务中表现更优，输出质量显著提升
- 适合需要高可靠性与多样性的工业级应用开发者参考
生成式预训练变换器（GPT）是文本生成的基础大语言模型（LLM），但单个LLM常产生不一致输出并带有偏见，难以全面反映语言模式。许多强大LLM为闭源，因数据隐私问题限制了行业应用。受文本生成中集成成功经验启发，如今代码生成领域也兴起对LLM集成技术的研究。本文综述了这些新兴的集成方法，旨在深化理解、推动研究并促进实际应用。我们将其分为七类：权重融合、知识融合、专家混合、奖励集成、输出集成、路由机制与级联结构，并分析各方法能力。研究发现，集成可有效提升多样性表示、输出质量和应用灵活性，有助于真实场景中的模型选型，并为扩展至多模态大模型奠定基础。
原文摘要 · Abstract (English)
Generative Pretrained Transformers (GPTs) are foundational Large Language Models (LLMs) for text generation. However, individual LLMs often produce inconsistent outputs and exhibit biases, limiting their representation of diverse language patterns. The closed-source nature of many powerful LLMs further restricts industry applications due to data privacy concerns. Inspired by successes in text generation, LLM ensemble techniques are now increasingly explored for code generation. This article reviews these emerging ensemble approaches to enhance understanding, encourage further research, and promote practical implementation in both text and code generation. We categorize LLM ensembles into seven main methods - weight merging, knowledge fusion, mixture-of-experts, reward ensemble, output ensemble, routing, and cascading - analyzing capabilities of those approaches. Our findings highlight key benefits such as improved diversity representation, enhanced output quality, and greater application flexibility. These insights aid model selection for real-world tasks and crucially, lay groundwork for extending ensemble strategies to multimodal LLMs.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。