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

Long-form Question Answering on AlpacaFact MixP (50% polluted)

81.7VeriScore F1@k

MIRAGE

44.374454.064763.75573.4453Jul 6, 2026
Updated 18d ago

Evaluation Results

MethodLinks
2026.07
81.7
2026.07
76.41
2026.07
72.34
2026.07
70.38
2026.07
69.97
2026.07
68.11
2026.07
67.43
2026.07
65.73
2026.07
65.71
2026.07
64.98
2026.07
63.57
2026.07
61.42
2026.07
61.12
2026.07
60.62
2026.07
60.53
2026.07
59.59
2026.07
59.13
2026.07
58.92
2026.07
57.32
2026.07
57.1
2026.07
55.34
2026.07
55.26
2026.07
55.15
2026.07
53.88
2026.07
53.61
2026.07
45.81