Exploiting LLMs for Automatic Hypothesis Assessment via a Logit-Based Calibrated Prior | Read Paper on Bytez