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

Language Modeling on WikiText-103 zero-shot (test)

12.76PPL

Megatron-LM GPT-2

10.26427.11243.9660.808Aug 13, 2021May 10, 2022Feb 5, 2023Nov 3, 2023Jul 30, 2024Apr 27, 2025Jan 23, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2021.08
12.76
2021.08
13.72
2021.08
13.88
2021.08
13.88
2021.08
13.89
2021.08
13.89
2021.08
14.14
2021.08
14.21
2021.08
14.76
2021.08
17.48
2021.08
19.31
2026.01
25.55
2021.08
26.03
2021.08
27.01
2021.08
27.06
2026.01
27.09
2021.08
27.15
2021.08
27.74
2021.08
27.77
2021.08
27.78
2021.08
28.09
2021.08
28.19
2026.01
29.29
2026.01
29.98
2026.01
31.39
2026.01
35.15
2026.01
35.66
2021.08
37.5
2026.01
37.98
2026.01
40.62
2026.01
41.6
2026.01
49.6
2026.01
50.86
2026.01
75.16