Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Factual Accuracy Estimation on Real-data Factual GPT-5.4 responses

64.39Original Accuracy

Logistic Regression

61.170562.7802564.3965.99975Jun 18, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
64.3958.72-5.6885
2026.06
64.3959.01-5.3885
2026.06
64.3963.4-0.9985
2026.06
64.3954.87-9.5285
2026.06
64.3964.39085
2026.06
64.3967.663.2795.6