Multiple-Choice Question Answering on STEM 4-Choice
0.3Delta AccuracyBatch Calibration (BC)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Batch Calibration (BC)Model=Llama3.1-8B2026.02 | 0.3 | — | — | -0.033 | |
| RBCorrModel=Llama3.1-8B2026.02 | 0.2 | — | — | -0.108 | |
| Batch Calibration (BC)Model=Llama3.1-8B-Instruct2026.02 | -0.2 | — | — | -0.033 | |
| RBCorrModel=Llama3.1-8B-Instruct2026.02 | -0.4 | — | — | -0.02 | |
| Contextual Calibration (CC)Model=Llama3.1-8B-Instruct2026.02 | -1.7 | — | — | 0.111 | |
| Contextual Calibration (CC)Model=Llama3.1-8B2026.02 | -10.5 | — | — | 0.37 | |
| BaselineModel=Llama3.1-8B2026.02 | — | 52.1 | 0.146 | — | |
| BaselineModel=Llama3.1-8B-Instruct2026.02 | — | 56.5 | 0.068 | — |