一作ICLR 2026Poster· Deep Learning / Algorithms
Dr.LLM: Dynamic Layer Routing in LLMs
Ahmed Heakl · Martin Gubri · Salman Khan · Sangdoo Yun · Seong Joon Oh
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。
Ahmed Heakl · Martin Gubri · Salman Khan · Sangdoo Yun · Seong Joon Oh
Ahmed Heakl · Gustavo Bertolo Stahl · Sarim Hashmi · Seung Hun Eddie Han · Mukul Ranjan · Arina Kharlamova · Salman Khan · Abdulrahman Mahmoud
Cross-architecture GPU code transpilation is essential for unlocking low-level hardware portability, yet no scalable solution exists. We introduce CASS, the first dataset and model suite for source- and assembly-level GPU translation (CUDA ↔ HIP, SASS ↔ RDNA3). CASS contains 60k verified host-dev...
Rania Elbadry · Sarfraz Ahmad · Ahmed Heakl · Dani Bouch · Momina Ahsan · Muhra AlMahri · Marwa Elsaid Khalil · Yuxia Wang · Salem Lahlou · Sophia Ananiadou · Veselin Stoyanov · Jimin Huang · Xueqing Peng · Preslav Nakov · Zhuohan Xie
English financial NLP has progressed rapidly through benchmarks for sentiment, document understanding, and financial question answering, while Arabic financial NLP remains comparatively under-explored despite strong practical demand for trustworthy finance and Islamic-finance assistants. We intro...
Cornelius Emde · Alexander Rubinstein · Anmol Goel · Ahmed Heakl · Sangdoo Yun · Seong Joon Oh · Martin Gubri
The rapid adoption of LLM-based agentic systems has produced a rich ecosystem of frameworks (smolagents, LangGraph, AutoGen, CAMEL, LlamaIndex, i.a.). Yet many existing benchmarks are model-centric: they fix the agentic setup and do not compare other system components. We argue that implementatio...