INDEX

Explanations

references to the act of lying

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

layout

-1.20

 layout

-1.02

 Layout

-0.92

Layout

-0.91

 layouts

-0.82

 LAYOUT

-0.63

layouts

-0.61

 Lawton

-0.58

PerformLayout

-0.58

LAYOUT

-0.58

POSITIVE LOGITS

lie

3.36

 lies

2.88

Lie

2.50

 Lies

2.38

LIE

2.33

Lie

2.28

Lies

2.25

 lying

2.23

 Lying

2.02

 lied

1.95

Activations Density 0.024%