INDEX

Explanations

references to self-awareness and a humorous approach to serious situations

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Negative Logits

shore

-0.07

 Lazar

-0.06

_lazy

-0.06

isy

-0.06

ply

-0.06

fik

-0.06

ogue

-0.06

illy

-0.06

_RST

-0.06

 PackageManager

-0.06

POSITIVE LOGITS

 seriousness

0.12

 Serious

0.11

 serious

0.11

serious

0.09

 seriously

0.08

 relaxed

0.07

 Relax

0.07

-ser

0.07

 Seriously

0.07

Seriously

0.07

Activations Density 0.013%

No Known Activations