INDEX

Explanations

a lot of

This neuron detects the common “a lot of” quantifier phrase indicating a large amount.

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

-1.86

 want

-1.80

 when

-1.70

who

-1.68

has

-1.68

 where

-1.67

my

-1.66

the

-1.64

if

-1.63

 than

-1.63

POSITIVE LOGITS

哢

1.74

риях

1.63

 kumaş

1.56

viele

1.52

 हूं

1.48

 bakter

1.45

🕉

1.43



1.43

Thanos

1.41

 procès

1.38

Activations Density 0.028%