INDEX

Explanations

assumptions; also

The neuron fires strongly on words that introduce or label theoretical premises—e.g. “assumption,” “assumptions,” “approximation,” or “axioms.”

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

！"

-1.50

 chaises

-1.42

 ihres

-1.40

ючи

-1.40

 brinde

-1.38

rieron

-1.38

。"

-1.36

💸

-1.33



-1.32

йдет

-1.30

POSITIVE LOGITS

↵↵

1.80

 Also

1.49

 también

1.48

by

1.46

 também

1.45

世上

1.42

Also

1.40

wanzig

1.33

 that

1.32

 neuesten

1.31

Activations Density 0.051%