INDEX

Explanations

say something

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zA

-0.85

LEncoder

-0.84

 tolerated

-0.78

 CallOverrides

-0.77

arResult

-0.77

 células

-0.76

 Música

-0.74

 RAILROAD

-0.74

rpt

-0.73

近平

-0.73

POSITIVE LOGITS

 instead

1.08

 real

1.07

真正的

0.98

 signified

0.95

 связан

0.90

instead

0.87

 REAL

0.86

Instead

0.86

 consiste

0.86

Truly

0.86

Activations Density 0.035%