INDEX

Explanations

I + verb

pronouns followed by will/can

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

してた

0.25

 udah

0.25

 been

0.24

 doesnt

0.24

 theyre

0.24

 wasnt

0.24

الم

0.24

<0xF3>

0.23

 Been

0.23

 theres

0.23

POSITIVE LOGITS

 Кстати

0.30

 übrigens

0.23

BTW

0.23

 кстати

0.23

再说

0.22

 strangely

0.22

 mischiev

0.22



0.22

 anyways

0.21

 అందుకే

0.21

Activations Density 0.289%