INDEX

Explanations

phrases that indicate expectations and responsibilities in a mentorship context

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

okoj

-0.09

ije

-0.08

oldem

-0.08

enou

-0.08

educt

-0.07

ivol

-0.07

Leaks

-0.07

akk

-0.07

Ð°Ð»ÐµÐ¶

-0.07

LOSS

-0.07

POSITIVE LOGITS

 mean

0.06

 Anders

0.05

 Cumberland

0.05

 sensitive

0.05

 perfectly

0.05

 Grove

0.05

 Hend

0.05

cul

0.05

rove

0.05

adians

0.05

Activations Density 0.026%