INDEX

Explanations

leader in providing

The neuron strongly activates on numeric tokens (including integers and decimals) representing quantitative metrics.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

atized

-1.87

 новом

-1.70

 vijf

-1.66

឵

-1.62

取决

-1.59

atization

-1.56

afety

-1.56

 ſol

-1.54

婠

-1.54

悫

-1.52

POSITIVE LOGITS

 providing

1.87

 what

1.81

 creating

1.70

Both

1.60

so

1.60

 making

1.59

 provide

1.52

 increased

1.52

 provides

1.48

of

1.48

Activations Density 0.106%