INDEX

Explanations

The neuron detects strongly positive or intense descriptive adjectives (e.g. “rich,” “active,” “vibrant,” “excellent,” “huge,” “strong”) conveying high degree or praise.

notable positive attributes

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

ător

-1.58

aneity

-1.55

ctomy

-1.54

 tendrá

-1.43

 podría

-1.43

荭

-1.41

 alcuni

-1.41

 potrebbe

-1.40

 wesentlichen

-1.40

ímos

-1.39

POSITIVE LOGITS

 range

1.55

1.38

 használ

1.37

 Decke

1.34

and

1.29

 variety

1.27

 mixture

1.24

 array

1.23

 collection

1.23

Activations Density 0.094%