INDEX

Explanations

code structure descriptions

The neuron fires on numeric measurements, especially tokens representing decimal numbers.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Apare

0.82

 Mening

0.80

 Appropri

0.72

اريخ

0.71

 असल्याचे

0.71

ע

0.71

נו

0.68

 publie

0.68

須

0.67

 ಎಂಬ

0.66

POSITIVE LOGITS

pathname

0.86

Voltage

0.81

Variant

0.80

globin

0.80

addLine

0.79

дные

0.76

NewLine

0.75

 novoProduto

0.75

ћа

0.75

Variable

0.74

Activations Density 0.001%