INDEX

Explanations

period followed by capital

This neuron detects statements of notable achievements or scale—lists of formal credentials, metrics, publications, or worldwide recognition.

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

or

-2.70

 then

-2.55

for

-2.34

 about

-2.20

all

-2.02

 those

-1.98

 sometimes

-1.94

 یا

-1.91

 или

-1.82

 when

-1.77

POSITIVE LOGITS

and

1.62

 plenti

1.52

 unab

1.48

 atyp

1.46

 renown

1.45

</b>

1.44

 undoubted

1.43

 consid

1.42

并且

1.38

举办

1.38

Activations Density 0.093%