Search papers, labs, and topics across Lattice.
The University of Texas at Austin
3
0
5
CalVerT boosts QA performance by equipping agents with calibrated self-confidence and grounding scores, reducing both erroneous confident answers and unnecessary information retrieval.
By selectively consulting the web only when uncertain, this fact-checking system achieves state-of-the-art results while maintaining interpretability, predictable cost, and conservative behavior.
LLMs in medical diagnosis are alarmingly prone to jumping to conclusions, often answering before seeing all the evidence, but strategically delaying the question and evidence presentation can boost accuracy by up to 62.6%.
