INDEX

Explanations

json log outputs

code function results

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

Exercises

0.40

ſon

0.40

Detectable

0.39

Diary

0.39

 ഞാന്‍

0.39

 सैक

0.38

CTOGRAM

0.38

 மலச்சிக்க

0.38

 सेकंड

0.38

^{(\

0.38

POSITIVE LOGITS

 provides

0.46

 pacing

0.46

 helps

0.45

 authoritative

0.45

 looks

0.44

 gives

0.44

==

0.43

 serves

0.43

 workplace

0.42

 cuáles

0.41

Activations Density 0.000%