How do humans and large language models (LLMs) represent sentences? Liu et al. design a rule inference task revealing that both humans and LLMs implicitly decompose sentences into tree-structured ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback