INDEX

Explanations

comparisons greater than or less than

The neuron detects numerical comparison expressions—specifically threshold tests like “less than X,” “greater than Y,” “above,” “below,” or “exceed” followed by a number.

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

bbean

-1.57

(!__

-1.46

ciating

-1.45

sellor

-1.43

NOWLEDGE

-1.38

but

-1.37

pagnol

-1.32

ifolium

-1.30

NSY

-1.30

opardy

-1.27

POSITIVE LOGITS

 nasceu

1.48

 empezaron

1.41

 allowing

1.34

ANY

1.29

 that

1.28

чему

1.27

 allowed

1.23

 для

1.23

starttime

1.22

 clearColor

1.21

Activations Density 0.038%