INDEX

Explanations

average or mean calculation

The neuron fires on numeric and quantitative tokens, identifying figures and statistics in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 اعزائي

0.67

MaxIntensity

0.65

可能性があります

0.63

 potentiel

0.60

 cris

0.59

℗

0.59

ล์

0.59

 apporter

0.57

 versione

0.57

ни

0.57

POSITIVE LOGITS

পড়

0.78

across

0.77

 平均

0.76

Joe

0.73

 арифмети

0.71

avg

0.70

Across

0.70

joe

0.69

annual

0.68

 aver

0.66

Activations Density 0.109%