INDEX

Explanations

years, numbers, and specific tokens

The neuron consistently activates on numeric tokens (years, page numbers, citation or section numbers) in the text.

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

 ویتامین

-1.73

 витами

-1.58

ittäin

-1.55

дры

-1.51

 Selasa

-1.50

פּ

-1.47

 десер

-1.41

ьере

-1.39

 dégustation

-1.38

Sú

-1.34

POSITIVE LOGITS

 бого

1.72

ſe

1.63

 Союза

1.59

 掃除

1.58

 cristianos

1.55

 мона

1.53

 IEEE

1.52

pæ

1.48

ſo

1.45

вѣ

1.44

Activations Density 0.002%