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

Character-Level Language Modeling on text8 (held-out 1M tokens)

2.3BPC

SHARP

Updated 23h ago

Evaluation Results

Method	Links
SHARP 2026.05		2.3
SHARP 2026.05		2.32
GRU 2026.05		2.38
Transformer 5M 2026.05		2.38
GRU 2026.05		2.41
Transformer 5M 2026.05		2.41
Transformer 10M 2026.05		2.41
Transformer 10M 2026.05		2.43
RNN 2026.05		2.45
RNN 2026.05		2.47
LSTM 2026.05		2.5
LSTM 2026.05		2.61
Clockwork RNN 2026.05		2.92
Clockwork RNN 2026.05		2.95