INDEX

Explanations

admit or deny

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

anka

-0.09

ceb

-0.07

 mens

-0.07

estr

-0.07

æĺ

-0.07

Iso

-0.07

uter

-0.07

 concl

-0.07

 Diss

-0.07

atto

-0.06

POSITIVE LOGITS

 admit

0.36

 deny

0.35

 admitting

0.34

 denial

0.34

 admits

0.31

 denies

0.31

 admitted

0.30

 denying

0.29

 denied

0.29

 admission

0.27

Activations Density 0.141%