List your company

Factuality Detection on TriviaQA-generated Fact Dataset topic-wise leave-one-out strategy 1.0

63.9Accuracy (billturnbull)SAPLMA
49.3453.1256.960.68May 22, 2025

Evaluation results

MethodLinks
SAPLMAModel=Llama 2-7b, Layer=16, Threshold Strategy=Optimal Threshold2025.0563.958.759.757.363.754.557.560.15661.555.361.958.354.956.6
SAPLMAModel=Llama 2-7b, Layer=20, Threshold Strategy=Optimal Threshold2025.0563.456.757.154.759.554.256.455.45556.95760.556.652.454.5
SAPLMAModel=Llama 2-7b, Layer=16, Threshold Strategy=Threshold = 0.52025.0562.859.461.758.963.553.858.265.457.361.157.161.857.556.656.5
SAPLMAModel=Llama 2-7b, Layer=20, Threshold Strategy=Threshold = 0.52025.0562.157.959.656.160.752.557.464.356.256.557.860.955.955.655.3
SAPLMAModel=Llama 2-7b, Layer=24, Threshold Strategy=Threshold = 0.52025.0561.856.857.552.358.852.156.56254.655.857.257.955.455.956.9
SAPLMAModel=Llama 2-7b, Layer=24, Threshold Strategy=Optimal Threshold2025.0561.855.855.952.160.15355.55553.255.556.157.455.953.656.1
SAPLMAModel=OPT-6.7b, Layer=24, Threshold Strategy=Threshold = 0.52025.0559.158.358.457.558.960.757.861.458.358.257.456.558.95557.8
SAPLMAModel=Llama 2-7b, Layer=28, Threshold Strategy=Optimal Threshold2025.0558.15556.552.358.75055.15852.554.956.155.155.45354.8
SAPLMAModel=Llama 2-7b, Layer=28, Threshold Strategy=Threshold = 0.52025.0557.656.458.154.758.551.855.862.753.655.757.556.555.455.456.3
SAPLMAModel=Llama 2-7b, Layer=last, Threshold Strategy=Threshold = 0.52025.0556.156.256.856.457.849.456.658.755.155.660.35656.853.657
SAPLMAModel=OPT-6.7b, Layer=28, Threshold Strategy=Threshold = 0.52025.0555.857.556.455.958.362.455.861.257.35756.556.359.155.157.3
SAPLMAModel=OPT-6.7b, Layer=28, Threshold Strategy=Optimal Threshold2025.0555.85656.65355.563.554.761.856.65554.955.658.247.655.2
SAPLMAModel=OPT-6.7b, Layer=20, Threshold Strategy=Threshold = 0.52025.0555.158.358.758.15959.65759.658.359.257.157.75958.758.4
SAPLMAModel=OPT-6.7b, Layer=last, Threshold Strategy=Threshold = 0.52025.0554.756.755.355.957.96255.960.65656.955.85658.453.654.8
SAPLMAModel=Llama 2-7b, Layer=last, Threshold Strategy=Optimal Threshold2025.0554.755.25654.857.650.855.854.353.754.160.256.356.450.656.7
SAPLMAModel=OPT-6.7b, Layer=24, Threshold Strategy=Optimal Threshold2025.0554.756.258.353.356.560.557.362.157.153.556.455.257.747.257.4
SAPLMAModel=OPT-6.7b, Layer=16, Threshold Strategy=Threshold = 0.52025.0553.755.557.25658.452.855.362.154.255.2545454.753.455.4
SAPLMAModel=OPT-6.7b, Layer=last, Threshold Strategy=Optimal Threshold2025.055355.356.254.455.963.254.561.855.153.75555.55744.953.5
SAPLMAModel=OPT-6.7b, Layer=20, Threshold Strategy=Optimal Threshold2025.055356.758.654.655.959.856.260.556.95856.457.557.351.157.9
SAPLMAModel=OPT-6.7b, Layer=16, Threshold Strategy=Optimal Threshold2025.0549.958.75854.154.855.355.161.952.150.853.953.753.647.257.1