INDEX

Explanations

hypothetical past conditions

The neuron strongly activates on occurrences of the auxiliary verb “have.”

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

 після

-1.75

 have

-1.72

卨

-1.45

しているので

-1.43

 humbly

-1.40

 graciously

-1.39

するので

-1.38

 !”

-1.33

 вже

-1.33

 bravely

-1.30

POSITIVE LOGITS

 också

1.49

'-

1.43

'$

1.40

 знаешь

1.36

聆

1.34

ons

1.33

-'

1.30

ook

1.30

ina

1.30

 sosteniendo

1.29

Activations Density 0.019%