Publication list sorted by year with filters for type and tags.
Taro Yamada, Alice Smith, Kenji Sato
We analyze failure modes of tool-augmented LLM agents in multilingual settings and introduce a reliability benchmark.
Mika Tanaka, Taro Yamada, Robert Lee
We reduce factual hallucinations by explicitly training evidence-grounded reasoning.
Taro Yamada, Emi Kobayashi
We propose a safe response optimization method combining human feedback and constrained decoding.
Haruto Watanabe, Taro Yamada, Yuki Ito
We evaluate continual pretraining and instruction tuning on low-resource domain corpora.