INDEX

Explanations

manipulation and deception

New Auto-Interp

Configuration

Prompts (Dashboard)

238,145 prompts, 512 tokens each

Dataset (Dashboard)

lmsys + oasst1

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 resp

0.53

 शौ

0.51

 upholding

0.50

 butch

0.50

 rigid

0.49

">

0.49

 regol

0.47

吼

0.46

 ochrony

0.46

">

0.46

POSITIVE LOGITS

 cunning

1.60

 deceptive

1.53

 deception

1.51

 manipulative

1.48

 manipulation

1.47

 deceive

1.46

 Manipulation

1.45

 deceiving

1.44

Manipulation

1.39

 manipulate

1.38

Activations Density 0.701%