INDEX

Explanations

class instantiation or definition

The neuron fires on numeric literals (number tokens) in the text.

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

Cuánt

-1.41

کفش

-1.34

presidente

-1.27

Cuánto

-1.26

 validamos

-1.23

klere

-1.23

atação

-1.22

ﾒ

-1.22

ญี่ป

-1.22

Эти

-1.22

POSITIVE LOGITS

1.67

1.45

1.39

1.38

to

1.34

1.33

1.30

1.28

our

1.27

1.23

Activations Density 0.001%