INDEX

Explanations

agree, warn, unfortunate

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

Abr

-0.11

/OR

-0.11

lus

-0.10

/TR

-0.10

 bliss

-0.10

åĬŁ

-0.10

,’’

-0.10

ENU

-0.10

 conce

-0.10

,’”

-0.09

POSITIVE LOGITS

 agree

0.12

agree

0.11

 warnings

0.11

 unfortunate

0.10

 unlucky

0.09

 recently

0.09

 notorious

0.09

 beware

0.09

 warned

0.09

 Agree

0.09

Activations Density 0.008%