INDEX

Explanations

phrases centered around moral and philosophical dilemmas

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 certainly

-0.07

ouser

-0.06

 Ð¾Ð±ÐµÑģÐ¿ÐµÑĩÐ¸Ð²Ð°

-0.06

af

-0.06

uf

-0.06

ãĥ¼ãĥŃ

-0.05

 definitely

-0.05

Ì

-0.05

åºĶ

-0.05

 Rath

-0.05

POSITIVE LOGITS

å¦ĤæŃ¤

0.09

ìĥī

0.07

è¿Ļä¹Ī

0.07

olly

0.07

habi

0.07

ÃŃÅ¾

0.07

 à¤ĩà¤¤à¤¨

0.07

_IOCTL

0.07

.:.

0.07

Ø§ØŃ

0.07

Activations Density 0.036%