INDEX

Explanations

numbers and letters

The neuron strongly activates on numeric tokens (digits or multi‐digit numbers).

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

˂

-1.48

 olje

-1.46

฿

-1.45

 that

-1.41

 évé

-1.38

 leche

-1.38

Ї

-1.35

of

-1.35

 saisi

-1.34

 vernis

-1.34

POSITIVE LOGITS

1.71

1.67

1.59

1.44

1.40

1.32

 alcune

1.22

 denselben

1.16

 wielu

1.16

 others

1.16

Activations Density 0.020%