INDEX

Explanations

if you're like

This neuron detects the informal comparative filler “like” when addressing or likening the reader to a group or trait.

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

who

-1.65

 findes

-1.60

💠

-1.52

 hulle

-1.48

 після

-1.45

 indien

-1.45

ecute

-1.42

ఒ

-1.39

🔆

-1.38

 toep

-1.37

POSITIVE LOGITS

 also

1.79

 также

1.68

 even

1.61

↵↵

1.56

by

1.52

1.46

for

1.45

 incluso

1.44

it

1.43

 such

1.43

Activations Density 0.176%