LLM增强LLMs：通过组合扩展能力

LLM Augmented LLMs: Expanding Capabilities through Composition

January 4, 2024

作者: Rachit Bansal, Bidisha Samanta, Siddharth Dalmia, Nitish Gupta, Shikhar Vashishth, Sriram Ganapathy, Abhishek Bapna, Prateek Jain, Partha Talukdar

cs.AI

摘要

具有数十亿参数的基础模型经过大规模数据训练，在各种领域展示了非平凡的技能。然而，由于它们的单olithic结构，要对其进行扩充或赋予新技能具有挑战性且昂贵。另一方面，由于它们的适应能力，正在训练这些模型的多个新实例以适应新领域和任务。在这项工作中，我们研究了有效和实用地将现有基础模型与更具体模型组合以实现新功能的问题。为此，我们提出了CALM -- 即增强语言模型的组合 -- 它引入了模型之间的交叉注意力，以组合它们的表示并实现新功能。CALM的显著特点包括：(i) 通过“重用”现有LLM以及少量额外参数和数据来扩展LLM在新任务上的规模，(ii) 保持现有模型权重不变，从而保留现有功能，以及(iii) 适用于不同领域和设置。我们阐述了通过在PaLM2-S上增加一个在低资源语言上训练的较小模型，使其在诸如翻译成英语和低资源语言的算术推理等任务上绝对改进高达13\%。同样，当在PaLM2-S中增加一个特定于代码的模型时，我们看到在代码生成和解释任务上相对于基础模型提高了40\% -- 与完全微调的对应模型相当。

English

Foundational models with billions of parameters which have been trained on large corpora of data have demonstrated non-trivial skills in a variety of domains. However, due to their monolithic structure, it is challenging and expensive to augment them or impart new skills. On the other hand, due to their adaptation abilities, several new instances of these models are being trained towards new domains and tasks. In this work, we study the problem of efficient and practical composition of existing foundation models with more specific models to enable newer capabilities. To this end, we propose CALM -- Composition to Augment Language Models -- which introduces cross-attention between models to compose their representations and enable new capabilities. Salient features of CALM are: (i) Scales up LLMs on new tasks by 're-using' existing LLMs along with a few additional parameters and data, (ii) Existing model weights are kept intact, and hence preserves existing capabilities, and (iii) Applies to diverse domains and settings. We illustrate that augmenting PaLM2-S with a smaller model trained on low-resource languages results in an absolute improvement of up to 13\% on tasks like translation into English and arithmetic reasoning for low-resource languages. Similarly, when PaLM2-S is augmented with a code-specific model, we see a relative improvement of 40\% over the base model for code generation and explanation tasks -- on-par with fully fine-tuned counterparts.