INDEX

Explanations

describing research objectives

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

for

-1.63

鼉

-1.55

 Accueil

-1.45

,—

-1.41

 focussing

-1.40

vns

-1.38

،

-1.38

 litros

-1.37

我好

-1.36

 lanka

-1.35

POSITIVE LOGITS

</em>

1.76

</strong>

1.58

</sup>

1.55

掼

1.51

1.49

<em>

1.48

1.44

 dizzy

1.42

1.35

 ві

1.34

Activations Density 0.049%