INDEX

Explanations

I suggest/recommend

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

 incluso

-1.02

 simpto

-0.92

 even

-0.88

 puto

-0.86

 jakarta

-0.84

 shouldBe

-0.81

atures

-0.81

 prezen

-0.80

ecer

-0.80

 kini

-0.79

POSITIVE LOGITS

 suggest

1.90

 suggesting

1.67

 suggests

1.54

 suggestion

1.43

建議

1.39

 would

1.37

建议

1.34

 assuming

1.31

suggest

1.27

 your

1.27

Activations Density 0.040%