Sketching Method Estimates Token Attribution With Far Fewer Evaluations
Researchers propose algorithm estimating leave-one-out attribution for all n prompt tokens using O(sqrt(log(1/epsilon))/epsilon) model evaluations, rather than n, under stability assumption. They also prove matching lower bound, showing method is asymptotically optimal. Why it matters Cheaper attri…
Read the full story at DEV Community — Machine Learning ↗
Timeline · 1 report
- 2026-11-15 07:02 · DEV Community — Machine Learning
Sketching Method Estimates Token Attribution With Far Fewer Evaluations