INDEX

Explanations

well

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

 renewed

-0.51

 cracks

-0.50

 decreases

-0.50

 greater

-0.48

 that

-0.48

 seized

-0.47

 differs

-0.47

}{@

-0.46

ONIA

-0.46

 huge

-0.46

POSITIVE LOGITS

PhysRevLett

0.56

 AssemblyCulture

0.56

 touristes

0.55

Encyklopedia

0.54

ConstraintMaker

0.53

----</

0.52

 Comptroller

0.51

schild

0.50

PhysRevD

0.49

RegressionTest

0.49

Activations Density 0.005%