Recurrent Alignment with Hard Attention for Hierarchical Text Rating

Chenxi Lin, Jiayu Ren, Guoxiu He +3

While large language models (LLMs) excel at understanding and generating plain text, they are not tailored to handle hierarchical text structures or directly predict task-specific…