CALM: LLM Augmented LLMs
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. How can we efficiently compose existing foundation models with more specific models to enable newer capabilities? CALM (Composition to Augment Language Models) 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. 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.
In this video, I talk about the following: How does CALM work? How does CALM perform for key-value arithmetic, low-resource language inclusivity and code understanding and generation?
For more details, please look at https://arxiv.org/pdf/2401.02412
Bansal, Rachit, Bidisha Samanta, Siddharth Dalmia, Nitish Gupta, Sriram Ganapathy, Abhishek Bapna, Prateek Jain, and Partha Talukdar. "LLM Augmented LLMs: Expanding Capabilities through Composition." In The Twelfth International Conference on Learning Representations. 2023.