INDEX

Explanations

makes sense or what i mean

makes sense

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

 eval

0.44

 coerced

0.43

 measurable

0.42

 coercion

0.42

 stringent

0.42

 waivers

0.42

 endangered

0.40

 eyesight

0.40

 clocked

0.40

 verifiable

0.40

POSITIVE LOGITS

 Brazil

0.60

 Brazilian

0.57

 Colombian

0.57

 Paragu

0.57

Brazilian

0.57

 Flamengo

0.57

 ব্রাজিল

0.55

巴西

0.55

 Suriname

0.54

 Luís

0.53

Activations Density 0.004%