INDEX

Explanations

phrases related to historical accuracy and verification

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

-0.07

 rhet

-0.06

 Arguments

-0.06

Î»Ïİ

-0.05

Arguments

-0.05

æ´ģ

-0.05

 Flores

-0.05

FAILURE

-0.05

ulos

-0.05

Î¸Î·

-0.05

POSITIVE LOGITS

 accuracy

0.21

 accurate

0.19

accuracy

0.19

 Accuracy

0.18

accur

0.18

Accuracy

0.16

 accur

0.16

 inaccurate

0.14

 correctness

0.14

 correct

0.14

Activations Density 0.063%