Madhav S Baidya

The best-performing text embedding model on benchmarks might not be the right choice for your specific retrieval system, and this framework reveals how to choose wisely.

LLMs struggle to recognize when they lack sufficient information to answer a question, but a surprisingly simple finetuning approach can teach them to abstain or ask for clarification, dramatically reducing hallucinations.

AI-generated text detectors that seem perfect in the lab fall apart in the real world, with no single method generalizing across domains or even different LLMs.