BlenderBot 3: a deployed conversational agent that continually learns to responsibly engage

TL;DR

BlenderBot 3 is a large-scale, internet-enabled conversational AI with long-term memory, designed for open-domain dialogue, safety, and continual learning, and has been publicly deployed for real-world interaction and research.

Contribution

We introduce BlenderBot 3, a 175B parameter dialogue model with internet access and long-term memory, and provide its deployment, safety mechanisms, and plans for continual learning from user interactions.

Findings

Outperforms previous open-domain dialogue agents in human evaluations

Successfully deployed for real-world user interactions

Plans for continual learning from deployment data

Abstract

We present BlenderBot 3, a 175B parameter dialogue model capable of open-domain conversation with access to the internet and a long-term memory, and having been trained on a large number of user defined tasks. We release both the model weights and code, and have also deployed the model on a public web page to interact with organic users. This technical report describes how the model was built (architecture, model and training scheme), and details of its deployment, including safety mechanisms. Human evaluations show its superiority to existing open-domain dialogue agents, including its predecessors (Roller et al., 2021; Komeili et al., 2022). Finally, we detail our plan for continual learning using the data collected from deployment, which will also be publicly released. The goal of this research program is thus to enable the community to study ever-improving responsible agents that learn through interaction.