INDEX

Explanations

verbs with following prepositions or conjunctions

The neuron selectively activates on common content verbs—especially transitive action verbs (e.g. “shown,” “gather,” “play,” “reduce,” “see,” “check”)—throughout the text.

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

たまに

-1.63

ét

-1.62

out

-1.55

︱

-1.41

 fortfarande

-1.41

 veta

-1.41

 antik

-1.35

aaaa

-1.35

+"/

-1.34

 autres

-1.32

POSITIVE LOGITS

be

2.36

 without

1.98

if

1.98

 with

1.69

 easily

1.66

 people

1.56

 dowol

1.51

 very

1.50

 various

1.48

 zajíma

1.45

Activations Density 2.918%