首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。
实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。
Masato Mita · Taiga Someya · Ryo Yoshida · Yohei Oseki
Large Language Models (LLMs) remain substantially less data-efficient than humans. Pre-pretraining (PPT) on synthetic languages has been proposed to close this gap, with prior work emphasizing highly expressive formal languages such as k -Shuffle Dyck. Inspired by the Language Acquisition Device...
Ryo Yoshida · Shinnosuke Isono · Taiga Someya · Yohei Oseki · Tatsuki Kuribayashi
Surprisal theory hypothesizes that the difficulty of human sentence processing increases linearly with surprisal, the negative log-probability of a word given its context. Computational psycholinguistics has tested this hypothesis using language models (LMs) as proxies for human prediction. While...
Tatsuki Kuribayashi · Alex Warstadt · Yohei Oseki · Ethan Gotlieb Wilcox
A recent study (CITATION) has shown that human sentence processing behavior, typically measured on syntactically unchallenging constructions, can be effectively modeled using surprisal from early layers of large language models (LLMs). This raises the question of whether such advantages of intern...