Goodtriever: Adaptive Toxicity Mitigation with Retrieval-augmented Models
This work addresses the need for adaptable toxicity mitigation techniques to handle evolving language in deployed models, though it is incremental in improving efficiency.
The paper tackles the problem of toxicity mitigation in text generation by introducing Goodtriever, a retrieval-augmented method that matches state-of-the-art performance while achieving a 43% relative latency reduction during inference.
Considerable effort has been dedicated to mitigating toxicity, but existing methods often require drastic modifications to model parameters or the use of computationally intensive auxiliary models. Furthermore, previous approaches have often neglected the crucial factor of language's evolving nature over time. In this work, we present a comprehensive perspective on toxicity mitigation that takes into account its changing nature. We introduce Goodtriever, a flexible methodology that matches the current state-of-the-art toxicity mitigation while achieving 43% relative latency reduction during inference and being more computationally efficient. By incorporating a retrieval-based approach at decoding time, Goodtriever enables toxicity-controlled text generation. Our research advocates for an increased focus on adaptable mitigation techniques, which better reflect the data drift models face when deployed in the wild. Code and data are available at https://github.com/for-ai/goodtriever.