INDEX

Explanations

format specifiers

The neuron activates on tokens containing digits, i.e. numeric strings.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 despite

-0.79

 each

-0.79

 except

-0.78

vén

-0.76

 exception

-0.75

keletal

-0.75

 separated

-0.75

 generally

-0.74

 owed

-0.74

ñez

-0.73

POSITIVE LOGITS

circular

0.93

DCC

0.88

 mại

0.84

 artinya

0.84

原始内容

0.83

 comrade

0.82

 गो

0.82

 knuckles

0.82

Geometry

0.81

 Emil

0.81

Activations Density 0.000%