INDEX

Explanations

comma

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 whether

-0.86

 Whether

-0.84

æĺ¯åĲ¦

-0.75

Whether

-0.75

whether

-0.72

èĥ½åĲ¦

-0.67

æĺ¯åĲ¦æľī

-0.65

 æĺ¯åĲ¦

-0.60

æĺ¯åĲ¦åŃĺåľ¨

-0.56

 WHETHER

-0.56

POSITIVE LOGITS

__["

0.33

 whats

0.30

__↵

0.27

(__

0.27

avigation

0.26

__↵↵

0.25

ä»Ģä¹Īåĳ¢

0.25

__).

0.25

__(↵

0.24

(___

0.24

Activations Density 0.981%