INDEX

Explanations

words introducing lists or explanations

contrasting common false beliefs

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ピアス

0.49

 polyn

0.48

 quella

0.47

ปลาย

0.46

ц

0.45

 पान

0.45

 平成

0.45

に関しては

0.45

 محاض

0.44

quels

0.43

POSITIVE LOGITS

arka

0.49

্যু

0.48

Detal

0.44

Dot

0.43

 Dirt

0.42

ंशिक

0.42

 absurd

0.41

retta

0.41

ამაშ

0.41

 ofta

0.40

Activations Density 0.002%