INDEX

Explanations

increasing degree

The neuron fires on degree‐modifying adverbs and comparatives (e.g. “more,” “further,” “rapidly”) that intensify or indicate an increase.

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

of

-1.71

 ندارم

-1.34

 icône

-1.32

 bête

-1.28

 eğiti

-1.23

 épaules

-1.20

</h2>

-1.18

 overpowering

-1.17

 lovable

-1.17

 mété

-1.17

POSITIVE LOGITS

ussure

1.45

 cantora

1.39

,,,

1.39

tividade

1.38

(\%)

1.34

way

1.34

ısında

1.33

 усилия

1.33

gencias

1.31

 encabez

1.31

Activations Density 0.033%