INDEX

Explanations

backwards, reversed, wrong way

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

Ἴ

-0.77

lehem

-0.76

 giờ

-0.75

 Riverside

-0.73

 Philippine

-0.71

braten

-0.71

 Szene

-0.71

Hentet

-0.71

 attrition

-0.71

ansas

-0.71

POSITIVE LOGITS

 backwards

5.47

 backward

4.63

backward

3.91

 Backward

3.48

Backward

3.33

 sideways

2.67

 reversed

2.55

 forwards

2.44

 reverse

2.33

 upside

2.16

Activations Density 0.047%