INDEX

Explanations

so

The neuron fires strongly on verbs (especially past‐tense or progressive forms), flagging action words in the text.

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

ícias

-1.25

return

-1.21

档案馆

-1.20

 irgend

-1.20

orderItem

-1.16

 worin

-1.15

 explicado

-1.13

uoš

-1.13

apunov

-1.12

鳢

-1.12

POSITIVE LOGITS

so

9.50

 так

4.94

如此

4.41

 such

4.25

 इत

4.16

这么

4.06

這麼

4.06

 così

4.03

那么

4.00

那麼

4.00

Activations Density 0.327%