INDEX

Explanations

multiplication and division by fractions

The neuron activates on individual digit tokens, i.e. it flags numeric characters (parts of numbers).

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

 նաև

-0.82

 contents

-0.79

ランド

-0.77

様に

-0.75

 futility

-0.74

 also

-0.73

mcm

-0.71

 콘텐츠

-0.70

 Autres

-0.70

cune

-0.70

POSITIVE LOGITS

half

0.80

Odd

0.79

~\

0.79

 Yong

0.78

Odd

0.77

一半

0.77

ographia

0.76

issä

0.76

Dav

0.76

 Cambridge

0.76

Activations Density 0.047%