AI Paper Atlas

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

Generalized Correctness Models: Learning Calibrated and Cross-Model Correctness Predictors from Historical Patterns

Hanqi Xiao, Vaidehi Patil, Hyunji Lee, Elias Stengel-Eskin, Mohit Bansal

ICML 2026 · Large language models › Language, safety and interpretability · approach: Uncertainty quantification for LLMs

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 a stronger or other model as predictor instead of the target itselfcorrectness model · MMLU answerlessaccuracy: Llama3-70B self .776 vs Qwen2.5-72B -> Llama3-70B .782No clear effect
train GCM on 8 models' histories instead of one SCM per modelQwen3-8B correctness model · MMLU (Llama3.1-8B target)accuracy: .792 -> .820Improved
train GCM on 8 models' histories instead of one SCM per modelQwen3-8B correctness model · MMLU (8 target models)AUROC: .825 -> .866Improved
add post-hoc spline calibration to GCMGCM · TriviaQA (GCM trained on MMLU)ECE: .105 -> .031Improved
apply GCM to held-out model Phi-3-miniGCM · MMLU (Phi-3-mini)accuracy: .787 -> .800Improved
apply GCM to held-out Qwen3-32BGCM · MMLU (Qwen3-32B)ECE: P(True) .074 -> .029Improved
remove target model name from GCM prompt (ablation)Answer-only GCM · MMLUaccuracy: .789 -> .763Hurt
add retrieved ICL examples to verbalized confidenceQwen3-32B predicting Gemma3-27B · MMLUaccuracy: +2.6% accuracy, -4.7% ECEImproved
add retrieved ICL examplesQwen3-8B verbalized confidence · MMLUaccuracyNo clear effect
train GCM without matched epochs (until overfitting) (ablation)GCM vs SCM · MMLU (Llama3.1-8B target)accuracy: SCM .822 vs GCM .845Improved
joint training on MMLU and TriviaQA (ablation)GCM · MMLU (Llama3.1-8B target)accuracy: .820 -> .827Improved
joint training on MMLU and TriviaQA (ablation)GCM · TriviaQA (Llama3.1-8B target)ECE: .029 -> .046Hurt

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