two followed by descriptions

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ara

1.67

or

1.59

1.57

1.51

1.47

uro

1.45

how

1.42

1.38

 Indian

1.37

oba

1.36

POSITIVE LOGITS

<unused1237>

2.45

<unused813>

2.43

湪

2.38

<unused846>

2.37

<unused9>

2.35

<unused1196>

2.34

𒄐

2.34

<unused1173>

2.34

<unused1465>

2.33

 যথেষ্ট

2.32

Activations Density 0.180%

No Known Activations