INDEX

Explanations

phrases related to maintenance and improvement

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

 eval

-0.06

omain

-0.06

ÑĢÐ¾ÑĪ

-0.06

 Rate

-0.06

Kov

-0.06

eldon

-0.06

akit

-0.06

ritten

-0.06

 Tire

-0.06

 chang

-0.06

POSITIVE LOGITS

 performance

0.10

performance

0.10

 PERFORMANCE

0.09

 Performance

0.09

Performance

0.08

 performances

0.08

-performance

0.08

æĢ§èĥ½

0.08

 performans

0.07

èĩ´

0.07

Activations Density 0.010%