INDEX

Explanations

NPR and National Public Radio

The neuron is strongly activated by numeric tokens (digits, numbers, and related numeric data) in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

Styles

-0.77

 turysty

-0.72

 дай

-0.71

 प्रमुख

-0.71

Resolved

-0.71

privile

-0.70

usza

-0.69

🅸

-0.69

 Waist

-0.69

PDG

-0.69

POSITIVE LOGITS

NPR

2.08

NPR

1.80

npr

1.47

Marketplace

0.88

 Qualität

0.87

Np

0.82

 Cosmopolitan

0.81

IER

0.75

Men

0.74

 dejan

0.73

Activations Density 0.053%