INDEX

Explanations

time units and durations

The main thing this neuron does is detect numeric tokens (numbers, dates, counts, or other digit sequences).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 различных

0.72

 Existem

0.68

 woven

0.67

 обеспечивает

0.66

 types

0.66

 various

0.66

 finns

0.60

 hierarchical

0.60

 provides

0.60

 tipos

0.59

POSITIVE LOGITS

weeks

1.16

晚上

1.13

month

1.13

February

1.10

早上

1.09

เดือน

1.09

 weeks

1.07

December

1.07

 tháng

1.06

 неделю

1.06

Activations Density 1.663%