Library-Like Behavior In Language Models is Enhanced by Self-Referencing Causal Cycles.

Munachiso Nwadike, Zangir Iklassov, Toluwani Aremu, Tatsuya Hiraoka, Benjamin Heinzerling, Velibor Bojkovic, Hilal AlQuabeh, Martin Takc, Kentaro Inui

Browse the full ACL paper archive.