List your company

Hallucination Detection on CoQA (dev)

77.8AUROC (Token Entropy)UA-CLM
51.997658.696365.39572.0937Dec 3, 2024

Evaluation results

MethodLinks
UA-CLMModel=Gemma-2B2024.1277.878.3773.5878.71
CLMModel=Gemma-2B2024.1277.2376.0672.9576.18
ULTModel=Gemma-2B2024.1270.9769.2165.471.62
Pre-trainedModel=Gemma-2B2024.1270.7370.8969.6276.35
UA-CLMModel=Llama-2-7B2024.1269.5573.9874.1377.41
UA-CLMModel=Llama-2-13B2024.1267.0172.5573.6376.94
ULTModel=Llama-2-13B2024.1263.2365.2368.8371.58
CLMModel=Llama-2-13B2024.1263.0263.4868.1569.1
CLMModel=Llama-2-7B2024.1262.5263.266.3568.89
CTModel=Llama-2-7B2024.1261.7565.7167.0672.92
Pre-trainedModel=Llama-2-13B2024.1260.2764.0466.7971.11
Pre-trainedModel=Llama-2-7B2024.1258.1363.2466.8674.67
ULTModel=Llama-2-7B2024.1257.959.1567.9364.95
CTModel=Llama-2-13B2024.1252.9955.9960.7269.58