INDEX

Explanations

The neuron strongly activates on standalone content nouns—especially abstract “topic” words like peace, mission, use, conception, and way.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

υτό

0.28

σιμοποι

0.27

quela

0.27

quele

0.26

ttä

0.26

 formulaire

0.25

orsanız

0.25

arankan

0.24

ামান

0.24

âtel

0.24

POSITIVE LOGITS

);

0.26

and

0.25

־

0.25

btw

0.24

 homers

0.24

0.23

).

0.23

Activations Density 0.000%