INDEX

Explanations

characters or symbols used in various contexts, indicating their presence or significance in the text

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

 Woody

-0.07

 Americ

-0.07

 Judy

-0.06

 Oprah

-0.06

hip

-0.06

 Cowboy

-0.06

 Whitney

-0.06

 Sinatra

-0.06

POP

-0.06

 Pulitzer

-0.06

POSITIVE LOGITS

Obs

0.13

Obs

0.11

Nec

0.10

 fucking

0.10

 Fucking

0.09

 fucks

0.09

 fuck

0.09

 fucked

0.09

 brutality

0.09

OBS

0.09

Activations Density 0.005%