INDEX

Explanations

quantifiable metrics and expressions of performance in various contexts

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

 ÐºÐ¾ÑģÑĤ

-0.07

ÛĮÙĨÚ¯

-0.07

_eof

-0.07

-ie

-0.06

pear

-0.06

enson

-0.06

ziej

-0.06

owitz

-0.06

MR

-0.06

POSITIVE LOGITS

 Lind

0.07

andbox

0.07

abant

0.06

ottage

0.06

eck

0.06

esk

0.06

eltas

0.06

Ïį

0.06

adel

0.06

à¹Įà¸Ł

0.06

Activations Density 0.002%