INDEX

Explanations

qualify for approval

The neuron fires on words that denote eligibility or approval status, such as “qualify,” “qualifying,” and “approved.”

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

to

-1.86

！"

-1.64

 '-':

-1.34

یمت

-1.29

nello

-1.28

 participan

-1.27

 если

-1.27

！(

-1.24

мым

-1.23

VENTORY

-1.23

POSITIVE LOGITS

travail

1.41

fauteuil

1.33

作った

1.30

ően

1.29

Remarque

1.23

</sup>

1.23

的一個

1.23

 défaite

1.23

,*

1.20

Exemple

1.20

Activations Density 0.011%