INDEX

Explanations

je prefix

This neuron fires on occurrences of the personal name “Jennifer” (and close spellings).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Cancell

0.59

 tér

0.56

isComposite

0.55

товано

0.54

eded

0.53

ები

0.52

ogén

0.52

 جال

0.52

 Canberra

0.52

了吧

0.52

POSITIVE LOGITS

 sort

0.60

İlk

0.59

პი

0.55

喷

0.55

 significant

0.54

妬

0.53

 painful

0.52

 ジェ

0.52

ἑ

0.52

 lingua

0.52

Activations Density 0.147%