INDEX

Explanations

preceding phrases describing scope

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 albeit

1.46

 tuttavia

1.42

 medida

1.41

 considerando

1.37

 unmittelbar

1.35

 alasan

1.35

 toutefois

1.34

 although

1.34

 secara

1.33

 direttamente

1.30

POSITIVE LOGITS

 वर्ल्ड

1.27

etc

1.27

UCLEAR

1.25

นู

1.24

 деше

1.22

めっちゃ

1.22

𝑋

1.17

etc

1.16

 وغیرہ

1.15

 overcrow

1.15

Activations Density 0.032%