INDEX

Explanations

watching crowd reactions

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

 willing

-0.11

 consenting

-0.10

sig

-0.09

 willingly

-0.09

 padded

-0.08

PN

-0.08

 crazy

-0.08

ARGIN

-0.08

 unprecedented

-0.08

gon

-0.08

POSITIVE LOGITS

 bystand

0.18

 observers

0.17

 watching

0.16

 nearby

0.15

 listeners

0.15

 audience

0.14

 viewers

0.14

 surrounding

0.14

 witnessing

0.13

 audiences

0.13

Activations Density 0.108%