INDEX

Explanations

dislike or hatred

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 reassuring

-0.12

 alarmed

-0.10

 worrying

-0.09

 Vulcan

-0.09

 reassure

-0.09

 blush

-0.09

ignet

-0.09

 innocence

-0.09

alar

-0.09

 normalization

-0.09

POSITIVE LOGITS

 hatred

0.61

 hate

0.61

 dislike

0.59

 hated

0.57

 hates

0.57

 Hate

0.50

 dislikes

0.48

 disliked

0.47

 desp

0.45

 hostility

0.40

Activations Density 0.398%

No Known Activations