Fetching the paper…

Mitigating Reversal Curse in Large Language Models via Semantic-aware Permutation Training · Around