Long-context Language Models Fail in Basic Retrieval Tasks Without Sufficient Reasoning Steps.

Yijiong Yu, Zhixiao Qi, Yongfeng Huang, Wei Wang, Weifeng Liu, Ran Chen, Ji Pei

Browse the full EMNLP paper archive.