Natural Language Inference on SNLI 3-Choice
11.7ΔAccContextual Calibration (CC)
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Contextual Calibration (CC)Model=Llama3.1-8B2026.02 | 11.7 | — | — | -0.029 | |
| RBCorrModel=Llama3.1-8B2026.02 | 5.4 | — | — | -0.222 | |
| Batch Calibration (BC)Model=Llama3.1-8B2026.02 | 4.3 | — | — | -0.143 | |
| RBCorrModel=Llama3.1-8B-Instruct2026.02 | 3.7 | — | — | -0.204 | |
| Batch Calibration (BC)Model=Llama3.1-8B-Instruct2026.02 | 2.2 | — | — | -0.082 | |
| Contextual Calibration (CC)Model=Llama3.1-8B-Instruct2026.02 | 1.2 | — | — | -0.095 | |
| BaselineModel=Llama3.1-8B2026.02 | — | 49 | 0.333 | — | |
| BaselineModel=Llama3.1-8B-Instruct2026.02 | — | 53.4 | 0.305 | — |