INDEX

Explanations

words said or described

The neuron fires on the main content words of a narrative—particularly nouns, verbs, and other semantically central 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

MANY

-1.45

vägen

-1.33

んですよ

-1.30

twelve

-1.24

sonder

-1.23

strange

-1.22

🏢

-1.21

℘

-1.18

 搞笑

-1.16

weird

-1.15

POSITIVE LOGITS

its

1.52

 another

1.41

 their

1.41

少しでも

1.28

As

1.27

aczy

1.27

 ihres

1.23

1.21

 encompasses

1.21

 again

1.21

Activations Density 0.022%