INDEX

Explanations

taught, learned, conditioned

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

 stalk

-0.13

 backstage

-0.11

stalk

-0.10

 redistrib

-0.10

 numbered

-0.10

 jealousy

-0.10

 unstoppable

-0.10

fbe

-0.10

 punct

-0.10

Wel

-0.10

POSITIVE LOGITS

 indo

0.25

 conditioning

0.24

 upbringing

0.23

 conditioned

0.22

 taught

0.21

 inherited

0.20

 learned

0.20

 Conditioning

0.20

 learnt

0.19

 trained

0.19

Activations Density 0.099%