INDEX

Explanations

phrases that denote quantities or sizes related to objects or concepts

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

ÐµÑģÐ¿

-0.07

stuff

-0.07

endale

-0.07

ÐµÐ»ÐµÐ¼

-0.07

pak

-0.07

utan

-0.07

NotAllowed

-0.06

ando

-0.06

 advice

-0.06

ĳľ

-0.06

POSITIVE LOGITS

few

0.12

 number

0.08

lot

0.08

Few

0.07

Few

0.07

few

0.07

ylan

0.07

edo

0.06

erable

0.06

isos

0.06

Activations Density 0.004%