INDEX

Explanations

the + noun

The neuron is specialized to detect the definite article “the.”

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

is

-1.71

of

-1.63

-1.51



-1.46

 alleging

-1.41

kmal

-1.38

</h1>

-1.36

 bett

-1.35

>");

-1.34

>")

-1.34

POSITIVE LOGITS

 trouva

1.73

 those

1.63

öpfer

1.60

ﾁ

1.54

ومه

1.47

Namun

1.44

Tentang

1.42

acakt

1.42

 vozid

1.41

NOUNC

1.40

Activations Density 0.026%