The training and inference of large language models (LLMs) are together a costly process that transports knowledge from raw data to meaningful computation. Inspired by the memory hierarchy of the human brain, we reduce this cost by equipping LLMs with explicit memory, a memory format cheaper than model parameters and text retrieval-augmented generation (RAG). Conceptually, with most of its knowledge externalized to explicit memories, the LLM can enjoy a smaller parameter size, training cost, and inference cost, all proportional to the amount of remaining "abstract knowledge". As a preliminary proof of concept, we train from scratch a 2. 4B LLM, which achieves better performance than much larger LLMs as well as RAG models, and maintains higher decoding speed than RAG. The model is named Memory³, since explicit memory is the third form of memory in LLMs after implicit memory (model parameters) and working memory (context key-values). We introduce a memory circuitry theory to support the externalization of knowledge, and present novel techniques including a memory sparsification mechanism that makes storage tractable and a two-stage pretraining scheme that facilitates memory formation.
Building similarity graph...
Analyzing shared references across papers
Loading...
Yang et al. (Mon,) studied this question.
www.synapsesocial.com/papers/68e61f51b6db6435875b1bc3 — DOI: https://doi.org/10.48550/arxiv.2407.01178
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:
Hongkang Yang
Zehao Lin
Wenjin Wang
Building similarity graph...
Analyzing shared references across papers
Loading...