INDEX

Explanations

here

let's see followed by observation

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 active

0.48

 hospitality

0.45

 aktif

0.45

 leadership

0.44

 gangsters

0.44

 passive

0.43

 midwifery

0.43

 sponsorship

0.43

 obstetric

0.42

 podcasts

0.41

POSITIVE LOGITS

 здесь

0.60

ここは

0.59

 акку

0.56

 aquí

0.55

 тут

0.54

 Здесь

0.54

 এখান

0.53

 여기가

0.53

 এখানে

0.53

 сюда

0.52

Activations Density 0.034%