INDEX

Explanations

record, need, made

technical or negative context

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 fighter

0.51

harp

0.47

 sincere

0.43

 optimistic

0.42

 Wisconsin

0.42

 night

0.42

 subsequent

0.41

 pencil

0.41

 uniq

0.40

 आवा

0.40

POSITIVE LOGITS

$&$-

0.54

 Fehler

0.53

werden

0.53

 errores

0.51

دد

0.51

('--

0.50

𒋾

0.49

<-

0.48

 ناخد

0.48

िड

0.48

Activations Density 0.000%