INDEX

Explanations

references to urban architecture and structures

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

oled

-0.06

onn

-0.06

åį

-0.05

ikel

-0.05

NI

-0.05

iti

-0.05

ROLL

-0.05

 correspondence

-0.05

rolls

-0.05

 traces

-0.05

POSITIVE LOGITS

 tower

0.19

 skys

0.19

 tallest

0.19

 towers

0.18

 tall

0.17

 taller

0.17

tower

0.15

 Tower

0.15

 Towers

0.15

 Tall

0.14

Activations Density 0.049%