INDEX

Explanations

language that addresses moral or ethical considerations regarding sin and redemption

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

panic

-0.08

ÑĹ

-0.07

rezent

-0.07

Ø§Ø±Ú©

-0.07

icus

-0.07

ãĥ¼ãĥľ

-0.07

rello

-0.07

åĲī

-0.07

.Buffer

-0.07

ÅĻez

-0.07

POSITIVE LOGITS

 instead

0.08

 treatment

0.07

 forg

0.07

 Instead

0.07

 compassion

0.07

 forgiving

0.07

 human

0.06

Instead

0.06

 forgiveness

0.06

 mercy

0.06

Activations Density 0.022%