INDEX

Explanations

The neuron is a “number detector” – it lights up on tokens carrying quantitative information (dates, counts, percentages, sports stats, mathematical numerals, etc.).

tied, equal, or balanced scores

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

ional

-0.84

Contin

-0.81

 esempio

-0.78

rician

-0.77

 الحكومة

-0.76

 версии

-0.75

 HERR

-0.75

 قل

-0.74

aller

-0.74

 Joshua

-0.74

POSITIVE LOGITS

 equal

1.41

 evenly

1.09

equal

1.06

 equals

1.04

Equal

1.03

 equality

1.03

 identical

1.02

EQUAL

1.00

 balanced

1.00

 Equal

0.99

Activations Density 0.135%