superscript numbers and symbols

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

loh

-0.11

 okhttp

-0.11

ondo

-0.10

igor

-0.10

uko

-0.10

istring

-0.10

 Engines

-0.10

rang

-0.10

rais

-0.09

.sd

-0.09

POSITIVE LOGITS

 Independent

0.08

 independently

0.08

aser

0.08

 Bien

0.07

 milit

0.07

 separately

0.07

 Stern

0.06

 conscious

0.06

Constraints

0.06

Activations Density 0.001%

No Known Activations