INDEX

Explanations

engineering

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 engineering

-1.30

 Prepared

-1.21

 engineer

-1.20

 prepared

-1.11

 wrong

-1.10

 Engineering

-1.09

 Engineer

-1.08

 engineered

-1.08

 engineers

-1.05

prepared

-1.02

POSITIVE LOGITS

 متعلقه

0.68

kloped

0.60

Amicalement

0.58

Hauptartikel

0.58

 createState

0.55

tagext

0.51

ness

0.48

']);

0.48

ing

0.48

."],

0.46

Activations Density 0.078%