/same-next-lang-token

Generate continuations with various (L)LMs and pick those that match into a dataset with higher worth for interpretability research

Apache License 2.0Apache-2.0

No issues in this repository yet.