INDEX

Explanations

if it were simple

The neuron activates on numeric tokens, i.e. any standalone numbers (digits or numerals) 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

and

-1.72

 två

-1.60

といった

-1.45

：……

-1.42

biggest

-1.41

 quando

-1.38

：“……”

-1.35

-1.34

ands

-1.29

level

-1.29

POSITIVE LOGITS

Would

1.72

Vgl

1.69

you

1.68

 serait

1.63

 would

1.54

 that

1.48

 Flä

1.47

Instead

1.45

 instead

1.45

 then

1.45

Activations Density 0.070%