INDEX

Explanations

make room for

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

 drowned

-0.11

 Wolff

-0.10

 exercised

-0.09

 franca

-0.09

apter

-0.09

éħį

-0.09

onia

-0.09

 Nass

-0.09

 Dias

-0.09

POSITIVE LOGITS

 supern

0.10

 Supern

0.10

Ko

0.09

 himself

0.09

Incoming

0.09

 clen

0.08

 Sting

0.08

fol

0.08

 herself

0.08

 Incoming

0.08

Activations Density 0.001%