INDEX

Explanations

research findings and reports

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

 strade

-0.81

 mine

-0.79

 Ezra

-0.77

特有

-0.77

 khá

-0.74

 Fairly

-0.74

 docile

-0.74

Terrain

-0.73

 compress

-0.73

 limestone

-0.72

POSITIVE LOGITS

 Live

1.00

Live

1.00

Passing

0.87

 خبرگزاری

0.80

passing

0.80

 live

0.80

cbo

0.78

Ḏ

0.74

SPACE

0.74

lives

0.73

Activations Density 0.016%