In-context Autoencoder for Context Compression in a Large Language Model

Open in new window