Discovering Interpretable Algorithms by Decompiling Transformers to RASP
Xinting Huang · Aleksandra Bakalova · Satwik Bhattamishra · William Merrill · Michael Hahn
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。
Xinting Huang · Aleksandra Bakalova · Satwik Bhattamishra · William Merrill · Michael Hahn
Blanka Kövér · Alexandra Butoi · Anej Svete · Michael Hahn · Ryan Cotterell
Satwik Bhattamishra · Kulin Shah · Michael Hahn · Varun Kanade
Yash Sarrof · Yupei Du · Katharina Stein · Alexander Koller · Sylvie Thiébaux · Michael Hahn
Entang Wang · Yiwei Wang · Aleksandra Bakalova · Michael Hahn
Doreen Osmelak · Yang Xu · Michael Hahn · Kate McCurdy
Languages vary widely in how meanings map to word forms. These mappings have been found to support efficient communication; however, this theory does not account for systematic relations within word forms. We examine how a restricted set of grammatical meanings (e.g. person, number) are expressed...
Tsan Tsai Chan · Varsha Suresh · Anisha Saha · Michael Hahn · Vera Demberg
Vision-language model (VLM) hallucination is commonly linked to imbalanced allocation of attention across input modalities: system, image and text. However, existing mitigation strategies tend towards an image-centric interpretation of these imbalances, often prioritising increased image attentio...