INDEX

Explanations

conditional phrases that express possibility or speculation

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

 Been

-0.09

Been

-0.08

 lately

-0.08

 recently

-0.07

 bisher

-0.07

 been

-0.07

aspers

-0.07

been

-0.07

ala

-0.06

qua

-0.06

POSITIVE LOGITS

âĢ¦âĢ¦ãĢĤ

0.07

Ã¼zel

0.06

enerator

0.06

adamente

0.06

dap

0.06

_handling

0.06

iod

0.06

izon

0.06

 fert

0.06

utom

0.06

Activations Density 0.068%