INDEX

Explanations

Stroup, field, ham Salih

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

Sanders

-0.10

 Woodward

-0.10

 Razor

-0.09

 overnight

-0.09

 resignation

-0.09

 destino

-0.09

 Sanders

-0.08

ilden

-0.08

 Montgomery

-0.08

 existential

-0.08

POSITIVE LOGITS

 Revelation

0.14

 lamb

0.11

 Lamb

0.11

 consum

0.11

 persec

0.10

 recept

0.10

 lake

0.10

 prow

0.10

 final

0.10

 Trib

0.10

Activations Density 0.154%