INDEX

Explanations

capitalized proper nouns

equations, Bloomberg, reduce

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

тный

0.43

 структуры

0.43

कई

0.42

 HOWEVER

0.41

 Aner

0.39

ITAL

0.39

 conceptual

0.39

しかし

0.39

EROS

0.37

 Einen

0.37

POSITIVE LOGITS

Bloomberg

0.48

ashvili

0.42

actyl

0.40

这家

0.39

 Bloomberg

0.38

 squeezing

0.38

MoveNext

0.38

 वस्तू

0.38

länder

0.37

 nullptr

0.37

Activations Density 0.026%