INDEX

Explanations

chaperone, cavalry

The neuron lights up on relatively long, multi-syllable nouns—especially uncommon or technical terms.

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

say

-2.77

-2.64

-2.31

-2.22

-2.17

ventude

-2.06

-1.95

Its

-1.88

mazioni

-1.86

POSITIVE LOGITS

ጵ

2.38

of

2.20

 tujuh

2.05

asanjo

2.03

usive

1.92

 Hanya

1.88

experiences

1.87



1.83

$~

1.83

りは

1.79

Activations Density 0.015%