INDEX

Explanations

reported, claimed, thought, revealed

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

 pretty

-0.24

 regs

-0.19

Âł

-0.18

 really

-0.17

pretty

-0.17

-ish

-0.17

 folks

-0.17

aka

-0.16

 whopping

-0.16

 goodies

-0.16

POSITIVE LOGITS

 “[

0.15

 âī¤

0.12

 âī¥

0.12

 âĪ

0.12

 reportedly

0.11

 ÃĹ

0.11

(Table

0.11

 (âĪ

0.11

 asynchronously

0.11

 subsequently

0.11

Activations Density 26.120%