INDEX

Explanations

anything or something

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

rios

-0.12

ousse

-0.11

/d

-0.11

rio

-0.11

ligt

-0.10

earer

-0.10

 something

-0.10

dio

-0.10

oor

-0.10

 terminator

-0.10

POSITIVE LOGITS

 else

0.26

_else

0.18

else

0.17

	else

0.15

Else

0.15

 Else

0.15

 ELSE

0.14

 elif

0.13

 resembling

0.12

elsen

0.11

Activations Density 0.071%

No Known Activations