INDEX

Explanations

headings and categories

The neuron primarily fires on longer, multi-syllable (i.e. longer character-length) tokens.

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

you

-1.66

-1.48

of

-1.32

was

-1.31

 since

-1.27

ของ

-1.26

 will

-1.24

 once

-1.24

 follows

-1.21

 unfortunately

-1.21

POSITIVE LOGITS

躥

1.43

kumpulan

1.39

—”

1.30

<h2>

1.27

 barnet

1.27

belea

1.26

 sehingga

1.26

还有

1.25

ତ

1.24

||

1.23

Activations Density 0.011%