RCS selects best-of-N LLM answers via weighted Fréchet mean of embeddings and radial distance ranking, outperforming majority voting and other baselines on seven benchmarks.
(11) Since the objective is strictly convex inz, this solution is unique
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Beyond Majority Voting: Efficient Best-Of-N with Radial Consensus Score
RCS selects best-of-N LLM answers via weighted Fréchet mean of embeddings and radial distance ranking, outperforming majority voting and other baselines on seven benchmarks.