INDEX

Explanations

escalators, reconsideration, switching

The neuron selectively activates on long, multi-syllable word fragments—especially those forming technical terms, proper names, or formal nouns.

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

 them

-1.53

 них

-1.45

 Mereka

-1.42

halts

-1.41

 vermel

-1.41

 nimmer

-1.40

 Médico

-1.34

鉨

-1.33

 vertus

-1.30

 oferta

-1.30

POSITIVE LOGITS

of

1.72

 that

1.50

—

1.35

by

1.34

に行ってきました

1.30

at

1.29

But

1.27

 untill

1.23

 Thursday

1.22

着一个

1.22

Activations Density 0.216%