A pool of language models can collaborate and improve collectively by learning from one another's responses. These interactions depend on the instructions us…
机构:华盛顿大学
来源:arXiv 2609.39076 | AI4Papers 论文推荐平台