Research Blog

Why Giving LLM’s a Memory Is Harder Than It Looks (2/3)
Why Giving LLM’s a Memory Is Harder Than It Looks (2/3)

How Language Models Remember, Forget, and Learn. In this three-part series we will explore the how and why of LLM memory. We will look at working memory, innate knowledge, and potential solutions to the problems we encounter. This second part will look at the “static brain” part of a model, how models get their innate knowledge and shed some light on the second kind of LLM amnesia.

Why Giving LLM’s a Memory Is Harder Than It Looks (1/3)
Why Giving LLM’s a Memory Is Harder Than It Looks (1/3)

How Language Models Remember, Forget, and Learn. In this three-part series we will explore the how and why of LLM memory. We will look at working memory, innate knowledge, and potential solutions to the problems we encounter. This first part will establish some basic terminology and shed some light on the first kind of LLM “forgetfulness”.

Newsletter

Schrijf je in op onze nieuwsbrief en ontvang meldingen over nieuwe research blogs!