作者位次 #4/5ICLR 2026Poster· Deep Learning / Attention Mechanisms
Probing Rotary Position Embeddings through Frequency Entropy
Yui Oka · Kentaro Hanafusa · Taku Hasegawa · Kyosuke Nishida · Kuniko Saito
可解释性OpenReview
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。
Yui Oka · Kentaro Hanafusa · Taku Hasegawa · Kyosuke Nishida · Kuniko Saito
Yui Oka · Itsumi Saito · Kyosuke Nishida · Kuniko Saito
Daiki Chijiwa · Taku Hasegawa · Kyosuke Nishida · Shin'ya Yamaguchi · Tomoya Ohba · Tamao Sakao · Susumu Takeuchi
Kazutoshi Shinoda · Kosuke Nishida · Kyosuke Nishida
Reward models (RMs) play a central role in aligning large language models (LLMs) with human preferences. However, RMs are often sensitive to spurious features such as response length. Existing inference-time approaches for mitigating these biases typically focus exclusively on response length, re...