RoPoLL: Robust Panel of LLM Judges

Researchers introduced RoPoLL, a robust panel of LLM judges that replaces standard consensus aggregation with a geometric median estimator, achieving unbounded bias resistance under contamination. In tests across 13 open-weight judges and three benchmarks, RoPoLL outperformed standard PoLL by up to 19% on biased corruption and enabled a 38B committee to beat a 675B single judge by 1.31x under 30% corruption.

arXiv:2606.30931v1 Announce Type: new Abstract: The LLM Jury, a Panel of LLM Evaluators PoLL reporting consensus scores, has become a practical alternative to single-judge LLM evaluation, yet its statistical behavior remains poorly understood. We formalize the LLM Jury under the Huber contamination model and show that PoLL incurs unbounded bias under any positive contamination, regardless of jury size, whenever a single judge fails in a biased, LLM-typical way mode collapse, sycophancy, safety refusal . Framing jury consensus as classical robust mean estimation, we propose RoPoLL Robust Panel of LLM-as-Judge , which preserves the PoLL panel but replaces the aggregation function with a robust mean estimator, instantiated with the geometric median GM : tuning-free, with the optimal finite-sample breakdown point 1/2. A finite-sample error bound and a matching information-theoretic minimax lower bound agree on the parametric rate sigma sqrt d/N and differ on the breakdown floor by a factor of sqrt d , a statistical-computational gap that polynomial-time RoPoLL pays relative to the intractable Tukey halfspace median. Across 13 open-weight judges 4B-675B , three reward-model benchmarks, and four corruption regimes at rates up to 50%, RoPoLL dominates PoLL on every biased corruption type: by about 19% on cross-dimensional attacks at matched compute, and by orders of magnitude on heavy-tailed Byzantine adversaries. A 3-judge RoPoLL committee at 38B beats Mistral-Large-3 675B by 1.31x on HelpSteer-2 under 30% bimodal-random corruption, an 18x parameter advantage at better accuracy; a Noisy-GT control confirms the premium is paid against biased contamination, not benign imprecision.