INDEX

Explanations

direction and its contexts

The neuron fires on terms that denote geometric orientation—especially words like “angle,” “direction,” “axis,” and “alignment.”

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

imantan

-2.11

 muerta

-1.68

 version

-1.66

וֹ

-1.64

habhar

-1.63

 выборе

-1.61

 relacionado

-1.59

zty

-1.56

rangian

-1.56

ASED

-1.56

POSITIVE LOGITS

of

1.76

the

1.52

好好的

1.50

probably

1.47

鸝

1.45

↵

1.45

presumably

1.43

those

1.43

ностью

1.42

 profond

1.41

Activations Density 0.019%