INDEX

Explanations

adverbs modifying actions or states

The neuron activates on adverbs (words ending in “-ly”) that intensify or qualify actions and descriptions.

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

 yüzük

-1.76

 Their

-1.57

」，

-1.47

-1.46

 rağmen

-1.45

 voient

-1.43

 halı

-1.43

 tuve

-1.40

 usos

-1.40

'},

-1.40

POSITIVE LOGITS

 both

1.66

 these

1.52

1.45

both

1.30

 этих

1.28

 любой

1.28

カナダ

1.28

any

1.27

 generando

1.27

<h2>

1.27

Activations Density 0.102%