1 paper
Christina Hahn, Shangbin Feng, Dean Light +3
A pool of language models can collaborate and improve collectively by learning from one another's responses. These interactions depend on the instructions used during training. Exi…