INDEX

Explanations

depicts or shows

The neuron fires on verbs indicating depiction or description (e.g. “depicts,” “depict,” “describes”).

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

."

-3.83

-3.02

↵↵

-2.69

☫

-2.69

ENT

-2.64

('

-2.59

-2.50

-2.42

},

-2.39

".

-2.39

POSITIVE LOGITS

菹

2.86

呿

2.78

 gewel

2.56

 Then

2.52

But

2.47

makeat

2.47

 verzame

2.44

 historische

2.42

 afectadas

2.42

The

2.39

Activations Density 0.005%