1 paper · 1 filter
Alan Dao, Dinh Bach Vu, Alex Nguyen +1
Small language models (SLMs) are inherently limited in knowledge-intensive tasks due to their constrained capacity. While test-time computation offers a path to enhanced performanc…