Edge 407: LLMs with Infininite Context Windows? Short-Term Memory and Autonomous Agents

The role of context windows in LLMs

Created Using Ideogram

In this Issue:

1. An overview of short term memory in autonomous agents.

2. Google’s Infini-Attention paper about transformers with nearly infinitive context windows.

3. The famous AutoGPT framework.

💡 ML Concept of the Day: Short-Term Memory and Autonomous Agents

In the previous issue of this series we discussed short and long term memory as the two fundamental types of memory in autonomous agents. In that taxonomy, short term memory refers to the information that is available in the context window. Usually, short term memory operates by mapping text inputs to a prompt template and relevant variables. Short term memory in LLMs is often equated to the role that RAM plays in computer systems as a rapid way to access information.