Unveiling Language-Specific Features in Large Language Models via Sparse Autoencoders.

Boyi Deng, Yu Wan, Baosong Yang, Yidan Zhang, Fuli Feng

Browse the full ACL paper archive.