List your company

Natural Language Inference on MNLI 3-Choice

9.9Delta AccuracyRBCorr
3.3485.0496.758.451Feb 12, 2026

Evaluation results

MethodLinks
RBCorrModel=Llama3.1-8B2026.029.9——-0.214
Batch Calibration (BC)Model=Llama3.1-8B2026.028.8——-0.14
Contextual Calibration (CC)Model=Llama3.1-8B2026.028.7——0.001
RBCorrModel=Llama3.1-8B-Instruct2026.024.7——-0.104
Contextual Calibration (CC)Model=Llama3.1-8B-Instruct2026.024.4——-0.017
Batch Calibration (BC)Model=Llama3.1-8B-Instruct2026.023.6——-0.076
BaselineModel=Llama3.1-8B2026.02—49.20.331—
BaselineModel=Llama3.1-8B-Instruct2026.02—59.70.221—