INDEX

Explanations

The neuron activates on uppercase two-letter tokens (like “MS”).

terms starting phrases

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

</th>

-1.44

 just

-1.39

or

-1.36

 without

-1.26

 first

-1.25

 lengthy

-1.24

 while

-1.23

 provide

-1.23

 their

-1.23

all

-1.23

POSITIVE LOGITS

癭

1.39

 deliciously

1.28

があるので

1.23

 insanely

1.22

 ridiculously

1.21

 delightfully

1.17

 классов

1.16

がこの

1.16

 DURING

1.16

having

1.16

Activations Density 0.000%