INDEX

Explanations

phenomena, principles, concepts, or themes

The neuron activates on high-level abstract or conceptual nouns (e.g. “phenomenon,” “principle,” “concept,” “technique,” “paradigm”) that introduce or label key theoretical ideas.

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

way

-1.55

 hittar

-1.40

and

-1.34

這些

-1.34

 verksamhet

-1.27

这些

-1.26

 Rückseite

-1.26

 undes

-1.26

 úrov

-1.25

 also

-1.24

POSITIVE LOGITS



1.66

by

1.55

احبه

1.46

vrigt

1.42

руем

1.32

 太郎

1.31

outro

1.30

沩

1.28

كامل

1.27

</sup>

1.27

Activations Density 0.086%