INDEX

Explanations

past participle followed by noun

This neuron activates on words ending in “-ed,” i.e. past-tense or past-participle forms.

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

⚌

-1.67

噲

-1.62

 الأخرى

-1.61

驁

-1.59



-1.56

ază

-1.55

foundland

-1.50

 cidade

-1.50

 bílé

-1.50

ኗ

-1.47

POSITIVE LOGITS

2.11

of

2.00

1.84

if

1.78

1.70

that

1.58

 predominantly

1.56

1.54

in

1.52

about

1.52

Activations Density 0.165%