INDEX

Explanations

protecting items with sleeves or covers

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

lle

-0.84

 drapes

-0.75

櫻

-0.74

طة

-0.73

siyon

-0.73

 paraffin

-0.71

rative

-0.71

 Clive

-0.71

lons

-0.71

illez

-0.70

POSITIVE LOGITS

 sleeved

1.04

 protective

1.01

sleeve

0.99

 sleeves

0.97

 protection

0.94

 sleeve

0.93

 protecting

0.91

 Protective

0.89

 bending

0.88

Sleeve

0.87

Activations Density 0.012%