Efficient Context Selection for Long-Context QA: No Tuning, No Iteration, Just Adaptive-$k$

Open in new window