INDEX

Explanations

numbers and fractions

This neuron does not reliably activate on any specific tokens or patterns.

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

 ſont

-1.66

 such

-1.63

↵

-1.58

摇了

-1.57

 três

-1.50

्रीन

-1.47

𝙖

-1.46

 fiables

-1.43

aggregate

-1.43

ates

-1.43

POSITIVE LOGITS

無論

1.51

南瓜

1.47

bidities

1.44

他們

1.40

韓

1.38

 This

1.37

也不

1.37

(≧

1.35

近年

1.34

 Afterward

1.34

Activations Density 0.006%