INDEX

Explanations

business and success strategies

The neuron responds to unusually long, multi-clause declarative sentences (often “tips” or explanatory statements) rather than individual words or short phrases.

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

-1.80

endimiento

-1.43

of

-1.40

“

-1.39

 получили

-1.38

upyter

-1.37

reicher

-1.32

Now

-1.30

 professora

-1.29

兵衛

-1.25

POSITIVE LOGITS

it

1.94

𝐍

1.84

新しい

1.64

 popoli

1.63

 élar

1.56

比特派

1.52

 такъ

1.51

 chré

1.51

 déchir

1.48

ᄐ

1.46

Activations Density 0.062%