Pre-training Limited Memory Language Models with Internal and External Knowledge

TL;DR

This paper proposes Limited Memory Language Models (LMLMs) that externalize factual knowledge to external databases during pre-training, enabling more transparent, editable, and verifiable knowledge while maintaining competitive performance.

Contribution

Introduction of LMLMs that externalize knowledge, allowing targeted lookups instead of memorization, improving transparency and editability of factual information.

Findings

LMLMs perform comparably to larger models on benchmarks.

LMLMs enable explicit knowledge editing.

External knowledge retrieval improves model transparency.

Abstract

Neural language models are black-boxes--both linguistic patterns and factual knowledge are distributed across billions of opaque parameters. This entangled encoding makes it difficult to reliably inspect, verify, or update specific facts. We introduce Limited Memory Language Models (LMLM), a new class of language models that externalizes factual knowledge to external database during pre-training rather than memorizing them. Our pre-training approach strategically masks externally retrieved factual values from the training loss, thereby teaching the model to perform targeted lookups rather than relying on memorization in model weights. Our experiments demonstrate that LMLMs achieve competitive performance compared to significantly larger LLMs on standard benchmarks, while offering the advantages of explicit, editable, and verifiable knowledge bases.