INDEX

Explanations

people and character names

This neuron activates on capitalized proper names (particularly person/character 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

it

-2.20

if

-2.03

-1.64

 rappresenta

-1.44

美味しいです

-1.38

 туры

-1.35

 interessanti

-1.35

 grazie

-1.35

how

-1.34

摂取

-1.33

POSITIVE LOGITS

ergies

1.60

 которы

1.40



1.39

">:

1.38

 می‌شوند

1.36

 assise

1.34

össä

1.32

","","

1.32

ícias

1.30

glPushMatrix

1.30

Activations Density 0.007%