LAMBADA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Lambada
89.7EM Accuracy
18
LAMBADA 8 reasoning tasks
55.61Accuracy
15
LAMBADA
11.39PPL
14
LAMBADA
73.2Accuracy
14
Lambada Standard
14.42PPL
13
LAMBADA (test)
66.51Accuracy
13
LAMBADA CONTROL (context)
65.6Accuracy
13
Lambada
21.6Perplexity
12
LAMBADA
74.9Accuracy
12
LAMBADA (control)
94Perplexity
12
Lambada
82.2Accuracy (Original)
11
LAMBADA
37.26Normalized Accuracy
10
Lambada
0.99MAUVE
9
LAMBADA OpenAI
59.3Accuracy
8
LAMBADA
62.3Accuracy
8
LAMBADA
53.79Coherence Score
8
LAMBADA
17.3Accuracy
7
LAMBADA (retention)
35.44PPL
7
LAMBADA
2Retention Drift (%)
7
LAMBADA Average over en, de, es, fr, it
18.2Zero-shot Accuracy
7
LAMBADA it
17.3Zero-shot Accuracy
7
LAMBADA fr
20.6Zero-shot Accuracy
7
LAMBADA es
16Accuracy (Zero-Shot)
7
LAMBADA de
18.3Zero-Shot Accuracy
7
LAMBADA en
18.8Zero-shot Acc
7