INDEX

Explanations

Hint to solution

The neuron fires on numeric tokens or number‐looking strings (e.g. digits, decimals, or multi‐digit values) 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

蟜

-1.69

satin

-1.50

 kommende

-1.40

鵙

-1.40

blouse

-1.37

 komplette

-1.35

vý

-1.29

fted

-1.28

 hadden

-1.27

 ПРЕ

-1.27

POSITIVE LOGITS

cę

1.54

 will

1.51

だが

1.45

ี้ยง

1.43

ufige

1.41

1.38

 provide

1.34

thirsty

1.34

 should

1.32

 continued

1.31

Activations Density 0.020%