INDEX

Explanations

value and its impacts

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

?>

-1.55

'}

-1.53

ються

-1.50

antaranya

-1.49

enfants

-1.48

konfirmasi

-1.47

[];

-1.46



-1.45

');

-1.43

玧

-1.43

POSITIVE LOGITS

 value

2.06

1.98

</h2>

1.88

value

1.83

</em>

1.69

 valuable

1.62

ve

1.48

1.47

––

1.47

あります

1.45

Activations Density 0.036%