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

Truthfulness on TruthfulQA (Reward, TR, IR, Div)

-1.8Reward

SEA

-6.3136-5.1418-3.97-2.7982May 26, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.05
-1.8899987
2025.05
-2.596710087
2025.05
-2.66769987
2025.05
-2.737410085
2025.05
-2.938010091
2025.05
-3679984
2025.05
-3.01779883
2025.05
-3.1729885
2025.05
-3.17759785
2025.05
-3.18649883
2025.05
-3.25739887
2025.05
-3.376610086
2025.05
-3.44799784
2025.05
-3.54869981
2025.05
-3.55689785
2025.05
-3.647810090
2025.05
-3.956710086
2025.05
-4.07789981
2025.05
-4.08769987
2025.05
-4.17649889
2025.05
-4.236210086
2025.05
-4.397710082
2025.05
-4.39769583
2025.05
-4.48709785
2025.05
-4.59649878
2025.05
-4.68738782
2025.05
-4.87729585
2025.05
-4.87729953
2025.05
-5.26669883
2025.05
-5.31559776
2025.05
-5.64759884
2025.05
-6.14599886