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

One Billion Word Benchmark

Benchmarks

Task NameDataset NameSOTA ResultTrend
Language ModelingOne Billion Word Benchmark (test)
Test Perplexity20.25
125
Language ModelingOne Billion Word Benchmark (val)
Perplexity26.6
11
Language ModelingOne Billion Word Benchmark
Perplexity35.1
10
Showing 3 of 3 rows