1 paper
Mo Yu, Tsz Ting Chung, Chulun Zhou +8
We introduce PRELUDE, a benchmark for evaluating long-context understanding through the task of determining whether a character's prequel story is consistent with the canonical nar…