INDEX

Explanations

words before specific roles or items

specific sequencing of words

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

їн

0.59

্রাজ

0.58

ід

0.55

 автомобі

0.53

arom

0.52

𝕖

0.50

ائلة

0.49

шеб

0.48

ನ್

0.47

Ꮂ

0.47

POSITIVE LOGITS

rai

0.46

 صحیح

0.46

 نمی‌

0.45

for

0.44

 یقین

0.44

 disintegration

0.44

spanning

0.43

 ihan

0.43

 particle

0.42

ising

0.42

Activations Density 0.001%