INDEX

Explanations

code initialization

The neuron strongly activates on function names (and calls) that perform initialization or setup (e.g. any “init” or “setup” routine).

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

various

-1.65

嵓

-1.41

 then

-1.41

$-\

-1.38

 différents

-1.37

 différentes

-1.37

 kemudian

-1.36

does

-1.35

 profili

-1.35

 verschillende

-1.34

POSITIVE LOGITS

new

1.70

 creador

1.33

</u>

1.30

 живе

1.23

 Only

1.23

𝘀

1.22

lastName

1.20

 rigorous

1.19

子さん

1.16

 anew

1.15

Activations Density 0.028%