INDEX

Explanations

essential role function

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

cursions

-2.03

雱

-1.88

鼱

-1.87

⡃

-1.83

 понравился

-1.80



-1.80

鷽

-1.79

 sœur

-1.79

꜆

-1.77

腘

-1.77

POSITIVE LOGITS

2.16

2.05

 essential

1.92

not

1.87

 Although

1.86

<strong>

1.80

now

1.79

at

1.78

re

1.73

1.71

Activations Density 0.039%