INDEX

Explanations

technical language

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hö

-0.50

興趣

-0.50

PONENT

-0.50

tirol

-0.46

 стаття

-0.46

tabular

-0.44

外人

-0.44

riculum

-0.42

τών

-0.42

 словарь

-0.42

POSITIVE LOGITS

AnchorStyles

0.79

0.75

pain

0.74

Pain

0.71

Datuak

0.70

 Pain

0.69

 pain

0.68

 PAIN

0.68

 useStyles

0.66

@"/

0.66

Activations Density 0.060%