INDEX

Explanations

casting spells and attacks

The neuron activates on numeric expressions—digits, decimal numbers, and related measurement or quantity tokens.

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

 tatlı

-1.72

 тебя

-1.50

Jahre

-1.46

？？？

-1.41

<0xF2>

-1.38

☑️

-1.36

émica

-1.33

ınızı

-1.29

 hubiera

-1.28

Jadi

-1.28

POSITIVE LOGITS

–

1.68

 pulsante

1.57

ússia

1.44

oléon

1.43

 Importantly

1.37

 모든

1.37

 festgestellt

1.36

 현재

1.33

âncias

1.32

“

1.31

Activations Density 0.011%