INDEX

Explanations

pauses, turns, what questions

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

hod

-0.07

çĲ¦

-0.07

 kabar

-0.07

æµĨ

-0.07

è®¤ç¼´åĩºèµĦ

-0.07

åħĨ

-0.07

resp

-0.06

å¥½å¥½çļĦ

-0.06

.ge

-0.06

ÑģÑĭ

-0.06

POSITIVE LOGITS

åıĪå¼Ģå§ĭ

0.09

è¯´åĪ°è¿ĻéĩĮ

0.09

omit

0.09

 again

0.08

åıĪéģĵ

0.08

again

0.08

 Again

0.08

åıĪæĥ³

0.08

åĨįæ¬¡

0.07

é¡¿äºĨé¡¿

0.07

Activations Density 0.012%