INDEX

Explanations

decided to give

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

åĽºçĦ¶

-0.07

po

-0.05

 curry

-0.05

ãĤ«æľĪ

-0.05

èĻ½çĦ¶æ²¡æľī

-0.05

é¦ĸåħĪ

-0.05

åĴĮäº¤æµģ

-0.05

åĩ¡

-0.05

rew

-0.05

åĴĮæľīåħ³

-0.05

POSITIVE LOGITS

alar

0.06

 Something

0.06

 Munic

0.06

Ø³ÙĦ

0.06

umuz

0.06

lect

0.06

lena

0.05

amel

0.05

ä¸įèĥ½ä½ľä¸º

0.05

Something

0.05

Activations Density 0.004%