INDEX

Explanations

going on, going to, going wrong

what's going on

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

ন্ম

0.42

FE

0.41

()%

0.37

compute

0.37

McC

0.37

 computation

0.36

 മണ്ണി

0.36

FG

0.35

 exacting

0.35

茣

0.35

POSITIVE LOGITS

 downhill

0.71

 unnoticed

0.70

 wrong

0.69

 viral

0.65

 nowhere

0.64

 overboard

0.62

 smoothly

0.62

 uphill

0.57

 astray

0.57

viral

0.54

Activations Density 0.082%