INDEX

Explanations

is the

The neuron consistently picks out copular or “to be” constructions (e.g. is, are, has been) that introduce key statements or definitions in academic writing.

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

 gesamte

-1.75

 then

-1.72

 Then

-1.69

 изменить

-1.63

 다른

-1.50

 nobody

-1.46

</h2>

-1.45

 wusste

-1.45

who

-1.44

whelmed

-1.38

POSITIVE LOGITS

 menemukan

1.47

as

1.42

 semanal

1.33

 unrelenting

1.32

tedì

1.31

 from

1.30

 meleg

1.30

 mesaj

1.30

 parfü

1.30

这家

1.27

Activations Density 0.100%