INDEX

Explanations

deciding to try things

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

你自己

-0.82

兼

-0.75

 причем

-0.75

anat

-0.75

 интерес

-0.74

нула

-0.72

Naturally

-0.71

AndEndTag

-0.71

selling

-0.71

 plut

-0.71

POSITIVE LOGITS

 regretted

1.23

 regrets

1.23

 glad

1.10

 regret

1.09

امل

1.05

 ahora

1.02

Turns

0.99

 arrep

0.95

 because

0.93

 ended

0.93

Activations Density 0.032%