Implicit meta-learning may lead language models to trust more reliable sources

1 code implementation • 23 Oct 2023 • Dmitrii Krasheninnikov, Egor Krasheninnikov, Bruno Mlodozeniec, Tegan Maharaj, David Krueger

Fine-tuning on this dataset leads to implicit meta-learning (IML): in further fine-tuning, the model updates to make more use of text that is tagged as useful.

In-Context Learning Meta-Learning