How I Study AI - Learn AI Papers & Lectures the Easy Way

Intermediate

Zixuan Huang, Xin Xia et al.Feb 9arXiv

Big AI reasoning models often keep thinking long after they already found the right answer, wasting time and tokens.

#SAGE#efficient reasoning#chain of thought

Intermediate

Amirhosein Ghasemabadi, Di NiuDec 23arXiv

Large language models often sound confident even when they are wrong, and existing ways to catch mistakes are slow or not very accurate.

#self-awareness#large language models#hidden states

Intermediate

Jun Zhang, Teng Wang et al.Dec 16arXiv

TimeLens studies how to teach AI not just what happens in a video, but exactly when it happens, which is called video temporal grounding (VTG).

#video temporal grounding#multimodal large language models#benchmark re-annotation

Papers3