INDEX

Explanations

instances of errors and mistakes

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

ierz

-0.07

loub

-0.07

åŀ

-0.07

 Lester

-0.06

ÑĢÐ¸Ð·

-0.06

mw

-0.06

rose

-0.06

 Flores

-0.06

UNUSED

-0.06

ENS

-0.06

POSITIVE LOGITS

/error

0.09

/errors

0.09

 mistakes

0.08

Made

0.07

iveness

0.07

 error

0.07

-made

0.07

 mistake

0.07

 rect

0.07

bote

0.07

Activations Density 0.017%