INDEX

Explanations

newspapereports

The neuron primarily activates on standalone numeric tokens (numbers, dates, ratings, etc.).

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

it

-2.44

We

-2.06

if

-2.03

</h4>

-1.90

 that

-1.73

 During

-1.73

-1.71

is

-1.67

in

-1.66

 because

-1.64

POSITIVE LOGITS

扨

1.88

ครื่อง

1.82

 Lunes

1.80



1.75

 eradic

1.69

 dát

1.67

ཥ

1.67

 kapture

1.63

OUTUBE

1.59

pielen

1.59

Activations Density 0.070%