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

Long-form Text Generation on FactScore

100Response Completeness

Greedy (Baseline)

33.54450.79768.0585.303Jan 3, 2026
Updated 3mo ago

Evaluation Results

MethodLinks
2026.01
10028.623.6
2026.01
74.239.853.5
2026.01
62.248.760.3
2026.01
50.542.864.4
2026.01
41.940.862.4
2026.01
41.249.968.1
2026.01
40.748.761.3
2026.01
37.545.763.8
2026.01
36.146.666.3