INDEX

Explanations

big/major/huge followed by consequence

The neuron strongly activates on the adjective “big,” i.e. it detects occurrences of the word “big.”

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

aneity

-1.66

 مناط

-1.62

᯾

-1.60

tront

-1.52

🩻

-1.50

炖

-1.45



-1.41

🫅

-1.41

yship

-1.41

appartamento

-1.41

POSITIVE LOGITS

ところに

1.73

為に

1.59

re

1.36

ことが

1.31

1.30

1.28

 dejando

1.27

1.26

ようだ

1.26

1.25

Activations Density 0.056%