INDEX

Explanations

impacted positively

toast, dim, hell, searches, letters

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

0.47

0.46

נ

0.46

0.45

ak

0.44

胤

0.43

hawk

0.43

0.42

λαν

0.41

boot

0.41

POSITIVE LOGITS

珽

0.57

 стату

0.57

 статус

0.52

 മു

0.51

 штат

0.51

 সঙ্গীতের

0.50

 профи

0.50

 पेंसिल

0.49

acuse

0.49

 Trat

0.48

Activations Density 0.000%