INDEX

Explanations

method called

The neuron fires on the literal token “called,” i.e. it finds occurrences of the word “called” in code or comments.

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

of

-1.86

其他的

-1.82

 aren

-1.80

 their

-1.79

他就

-1.59

 دو

-1.55

-1.52

を作る

-1.51

enerys

-1.50

 decorativa

-1.47

POSITIVE LOGITS

in

2.11

𓋼

1.95

いぐるみ

1.82

 Japão

1.80

 hardest

1.76

 stihl

1.74

 Senhora

1.74

 These

1.72

还不错

1.71

在這

1.69

Activations Density 0.014%