Paper page - Learning, Fast and Slow: Towards LLMs That Adapt Continually
…after training on one task, FST trained models adapt more effectively to a subsequent task than parameter-only trained models. In continual learning scenarios, where task domains change on the fly, FST…