1 paper · 1 filter
Haolang Lu, Hongrui Peng, WeiYe Fu +5
Failures in large language models (LLMs) are often analyzed from a behavioral perspective, where incorrect outputs in factual question answering are commonly associated with missin…