作者位次 #2/3NeurIPS 2026Poster
Evaluating Whether LLMs Can Reliably Connect the DOTs?
Eftekhar Hossain · John Salvador · Santu Karmaker
评测与基准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。
Eftekhar Hossain · John Salvador · Santu Karmaker
Md Mahadi Hassan Sibat · John Salvador · Akond Rahman · Santu Karmaker
LLMs show promise in code generation, yet their effectiveness for IT automation tasks, particularly for tools like Ansible, remains understudied. Existing benchmarks rely primarily on synthetic tasks that fail to capture the needs of practitioners who use IT automation tools. We present ExITBench...