INDEX

Explanations

normalized, adjusted, basis

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

</u>

-1.88

鐿

-1.53

}));

-1.50

');

-1.50

Especificación

-1.47

 alcune

-1.46

://$

-1.46

who

-1.45

ifero

-1.45

 noten

-1.43

POSITIVE LOGITS

1.87

1.77

at

1.66

1.43

1.37

1.30

ll

1.29

1.27

1.26

 their

1.25

Activations Density 0.070%