INDEX

Explanations

outcrop, flyer, away, issue

The neuron strongly activates on unusually long tokens (i.e. words or hyphenated/prefixed terms of many characters).

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

<h2>

-1.81

Both

-1.73

When

-1.71

Sometimes

-1.68

They

-1.63

During

-1.60

Maybe

-1.59

Usually

-1.56

Quite

-1.54

Two

-1.50

POSITIVE LOGITS

of

1.41

 hôtes

1.34

who

1.34

for

1.34

 senza

1.30

 ennemis

1.30

 organes

1.26

 zoals

1.25

 mojej

1.24

厛

1.23

Activations Density 0.219%