INDEX

Explanations

word endings

The main thing this neuron does is detect uncommon, multisyllabic English words (i.e. rare or sophisticated vocabulary).

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

 surrounding

-1.52

you

-1.49

ヌード

-1.48

are

-1.47

トイ

-1.42

 somewhat

-1.38

美味し

-1.35

埽

-1.35

などで

-1.34

実感

-1.34

POSITIVE LOGITS

 Septembre

1.40

+</

1.37

}~

1.34

?</

1.31

💞

1.30

kā

1.30

zelfde

1.29

...");

1.27

 cercle

1.27

izient

1.27

Activations Density 0.150%