How To Scale Transformers’ Memory up to 262K Tokens With a Minor Change?

Extending Transformers by memorizing up to 262K tokens

This article is a fabulous attempt to leverage language models in memorizing information by transformers with the least required effort. The point is that…