INDEX

Explanations

posted on

This neuron fires on proper names, especially personal names.

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

-1.70

\\

-1.66

-1.58

-1.52

-1.47

",

-1.46

ldots

-1.43

.'

-1.37

},

-1.36

-1.34

POSITIVE LOGITS

 撮っ

1.52

缫

1.45

 nantinya

1.40

urm

1.34

 boî

1.33

 最終

1.31

 comprende

1.30

 ビキニ

1.28

 véhicules

1.27

 munc

1.27

Activations Density 0.009%