INDEX

Explanations

whether or not

This neuron fires on Q&A–style phrasing, especially question marks and second-person pronouns (e.g. “you,” “do,” “or,” “?”).

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

 uniquement

-1.37

 sinta

-1.29

 mercen

-1.28

妩

-1.25

 فقط

-1.24

 profiter

-1.24

 terce

-1.23

 agir

-1.23

 herbes

-1.20

 applau

-1.19

POSITIVE LOGITS

if

1.92

 might

1.73

for

1.67

may

1.61

 whether

1.56

 menentukan

1.55

 Depending

1.48

 different

1.42

 consider

1.42

可能

1.41

Activations Density 0.025%