AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Who can we trust? LLM-as-a-jury for Comparative Assessment

Mengjie Qian, Guangzhi Sun, Mark Gales, Kate Knill

ICML 2026 · Large language models › Language, safety and interpretability · approach: Uncertainty quantification for LLMs · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add judge-specific discriminator to soft BT (BT-σ)soft BT jury aggregation · SummEval COHSpearman correlation: 53.94 -> 57.38Improved
add judge-specific discriminator to soft BT (BT-σ)soft BT jury aggregation · SummEval ALLSpearman correlation: 49.40 -> 50.50Improved
add judge-specific discriminator to soft BT (BT-σ)soft BT jury aggregation · Topical-Chat CNTSpearman correlation: 53.87 -> 56.30Improved
BT-σ instead of supervised temperature scaling (Temp-BT)jury aggregation · SummEval and Topical-ChatSpearman correlation: SummEval COH 56.21 -> 57.38Mixed
use aspect-dependent discriminators (BT-σ-asp) (ablation)BT-σ · SummEval and Topical-ChatSpearman correlation: SummEval COH 57.38 -> 57.36No clear effect
apply discriminators to hard comparisons (hard BT-σ) when inconsistency is high (ablation)hard BT · Topical-Chat ENGSpearman correlation: 62.57 -> 67.36Improved
hard BT-σ instead of soft BT-σ under moderate inconsistency (ablation)BT-σ · SummEval COHSpearman correlation: 57.38 -> 53.02Hurt
use soft instead of hard BT for individual inconsistent judgessingle LLM judge · SummEval COH (Qwen2.5-7B-Instruct)Spearman correlation: 51.42 -> 44.79Hurt
use soft BT instead of hard BT after aggregating multiple judgesjury aggregation · SummEval ALLSpearman correlation: 47.34 -> 49.40Improved
fit a Bradley-Terry model instead of averaging win probabilitiesAvg-Prob · SummEval ALLSpearman correlation: 45.15 -> 49.40Improved
soft BT instead of hard BT on Topical-Chat ENGjury aggregation · Topical-Chat ENGSpearman correlation: 62.57 -> 61.87Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy