INDEX

Explanations

this is where

The neuron strongly activates on occurrences of the word “where.”

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

in

-1.68

");

-1.60

 This

-1.60

as

-1.56

</td>

-1.54

chauung

-1.51

 disposizione

-1.49

has

-1.41

by

-1.38

เพราะ

-1.38

POSITIVE LOGITS

沨

1.97

 ofere

1.73

色列

1.59

他和

1.55

 verlangen

1.55

begins

1.55

 brachten

1.55

 culmin

1.53

starts

1.52

rii

1.52

Activations Density 0.010%