INDEX

Explanations

phrases after certain words to continue description

The neuron is specialized to detect numeric tokens—especially numbers with decimal points.

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

தல்

-1.52

ordnungen

-1.38

 alltid

-1.34

 altijd

-1.34

 länge

-1.31

klassi

-1.30

明显的

-1.29

｀*)

-1.27

klart

-1.27



-1.27

POSITIVE LOGITS

to

2.06

its

1.81

 that

1.71

if

1.68

1.52

 entweder

1.40

at

1.39

 here

1.38

1.37

 being

1.32

Activations Density 0.164%