INDEX

Explanations

and and contamination caused

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

veille

-0.07

locator

-0.07

åı¯ä½¿çĶ¨

-0.07

umÄģ

-0.07

çļĦç²¾

-0.07

ety

-0.07

ANIA

-0.07

isien

-0.07

cheiden

-0.07

åı¯å¥¹

-0.07

POSITIVE LOGITS

artic

0.07

ä¸Ģä¼ļ

0.07

 consequently

0.07

çĲª

0.07

×Ľ×Ł

0.07

Åĳt

0.07

 ëĤĺìĦľ

0.07

 importantly

0.07

arin

0.07

åĴ

0.06

Activations Density 0.756%