INDEX

Explanations

United States Court of Appeals

This neuron activates on the word “of.”

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

-1.77

➀

-1.61

年は

-1.43

 capteur

-1.40

when

-1.37

時には

-1.36

 oiseau

-1.34

Without

-1.34

-1.31

 wusste

-1.30

POSITIVE LOGITS

 ktorí

1.88

 ktorá

1.88

↵↵

1.71

 kteří

1.66

 která

1.54

1.53

 ktorý

1.52

 meinem

1.51

⍢

1.50

 แต่

1.49

Activations Density 0.001%