INDEX

Explanations

database table names

The neuron fires on standalone technical identifiers or section/file names (e.g. directory or menu labels) in the text.

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

Abl

0.60

the

0.59

 substituting

0.55

 immisc

0.55

 Explo

0.51

पासून

0.50

 Flam

0.50

 putting

0.47

 peroxide

0.47

 exploiting

0.47

POSITIVE LOGITS

DataManager

0.64

gegevens

0.62

/${

0.61

Info

0.59

 каждого

0.58

 அவர்களின்

0.57

 вопросов

0.56

 ค่ะ

0.56

 ซึ่ง

0.56

/"

0.55

Activations Density 0.000%