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

Factual Question Answering on AURATown environment-grounded queries N=50, 3 seeds 1.0

76.6Factual Accuracy

Fixed-Probe

4.21623.00841.860.592Jun 4, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
76.68563.2100.006
76.44.6458.75.2-0.004
2026.06
69.61.475.34.3100.098
2026.06
64066.73.90.001-
2026.06
60.3075.32.20.00050.06
55-51.360.10.033
450701.80.03210
2026.06
17.42.1555.920.6-10
7093.32.3-10