INDEX

Explanations

they are

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 laquelle

1.74

1.68

 koje

1.67

 следующих

1.62

 яких

1.57

ً

1.57

1.54

1.53

InstanceOf

1.52

نان

1.50

POSITIVE LOGITS

<unused2223>

2.75

eding

2.64

<unused1157>

2.53

<unused2138>

2.35

<unused2222>

2.30

pherds

2.25

duled

2.20

 hesitated

2.16

<unused2003>

2.16

pherd

2.14

Activations Density 0.569%