INDEX

Explanations

question or section markers

This neuron fires on philosophical or spiritual terms, especially “evolution” and “consciousness.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 réinitial

-1.59

 récol

-1.51

nữ

-1.48

 éte

-1.45

 konfer

-1.41

Indien

-1.40

 Demokrat

-1.38

 goû

-1.38

 publik

-1.36

 délais

-1.34

POSITIVE LOGITS

of

1.58

 which

1.55

as

1.51

or

1.47

 from

1.45

at

1.43

 does

1.41

 still

1.37

 with

1.35

 even

1.31

Activations Density 0.000%