INDEX

Explanations

here announcing presence

The neuron fires strongly on the word “here” (especially in “is here”) in headline‐style announcements to mark the arrival or release of something.

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

Öl

-1.45

—

-1.37

wiſe

-1.32

٢

-1.28

isesti

-1.27

 bleiben

-1.27

selt

-1.26

ｒ

-1.25

 двох

-1.25

嚶

-1.24

POSITIVE LOGITS

–

1.57

 Such

1.56

</b>

1.52

Our

1.48

 Another

1.45

 lesquelles

1.45

him

1.44

1.42

to

1.39

Activations Density 0.013%