AI Paper Atlas

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

SCOPE: Selective Conformal Optimized Pairwise LLM Judging

Sher Badshah, Ali Emami, Hassan Sajjad

ICML 2026 · Large language models › Language, safety and interpretability · approach: Statistical frameworks for LLM evaluation · 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
use BPE (average forward/reverse probabilities, binary entropy) instead of Swap-and-Aggregate discrete voting (ablation)bidirectional LLM judge · MT-Bench (Qwen2.5-14B)AUROC: 0.679 -> 0.777Improved
use BPE instead of Swap-and-Aggregate (ablation)bidirectional LLM judge · RewardBench (Qwen2.5-7B)AUROC: 0.673 -> 0.744Improved
use BPE instead of Predictive ProbabilityLLM judge · MT-Bench (Qwen2.5-7B)ECE: 0.239 -> 0.143Improved
use BPE instead of Verbalized ConfidenceLLM judge · MT-Bench (Qwen2.5-7B)AUROC: 0.497 -> 0.685Improved
use BPE instead of Verbalized ConfidenceLLM judge · MT-Bench (Qwen2.5-14B)ECE: 0.114 -> 0.174Hurt
use BPE instead of Simulated AnnotatorsLLM judge · MT-Bench (Qwen2.5-7B)AUROC: 0.59 -> 0.69Improved
use BPE instead of Simulated AnnotatorsLLM judge · MT-Bench (Qwen2.5-14B)AUROC: 0.60 -> 0.78Improved
use BPE instead of Simulated AnnotatorsLLM judge · RewardBench (Qwen2.5-7B)ECE: 0.103 -> 0.104No clear effect
use SCOPE conformal thresholding instead of naive calibrationselective LLM judging · MT-Bench (Qwen-7B, alpha=0.10)empirical risk: 0.116 -> 0.097Improved
use SCOPE instead of naive calibrationselective LLM judging · MT-Bench (Qwen-14B, alpha=0.10)coverage: 0.566 -> 0.504Hurt
use SCOPE instead of naive calibrationselective LLM judging · MT-Bench (Qwen-7B, alpha=0.10)coverage: 0.102 -> 0.246Improved
use SCOPE instead of heuristic thresholdingselective LLM judging · MT-Bench (Qwen-7B, alpha=0.10)empirical risk: 0.251 -> 0.097Improved

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