INDEX

Explanations

introduces a consequence

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

 zwei

-2.77

</i>

-2.73

虓

-2.59

 sechs

-2.58

 sufri

-2.56

 aña

-2.53

</b>

-2.53

</h1>

-2.50

 cinco

-2.50

</th>

-2.47

POSITIVE LOGITS

！！”

3.31

墬

3.25

鲧

3.11

</em>

2.86

 δεν

2.83

！！！”

2.81

 ,‬

2.80

‪.‬‬

2.73

～

2.69

鹇

2.66

Activations Density 0.002%