INDEX

Explanations

validity context

The neuron activates on numeric measurements and their context—especially decimal or integer values (e.g. “41.5,” “24.125,” “valid 156,” “for 164,” etc.).

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

 diversos

-1.55

↵

-1.52

-1.51

 mientras

-1.43

 diverses

-1.41

-1.40

</strong>

-1.33

 zwei

-1.31

type

-1.29

 mène

-1.27

POSITIVE LOGITS

for

1.96

on

1.72

Even

1.37

(),

1.27

 विश्वसनीयता

1.26

Probably

1.25

Regardless

1.24

=“

1.23

={

1.23

 Organisationen

1.20

Activations Density 0.013%