INDEX

Explanations

emotional states and responses

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

';

-1.87

UCIÓN

-1.83

);

-1.77

噢

-1.69

 Please

-1.59

asanjo

-1.59

to

-1.58

";

-1.58

 were

-1.56

];

-1.55

POSITIVE LOGITS

↵↵

1.77

 kwamba

1.66

「

1.60

1.59

 stoppen

1.56

 Emotions

1.54

もちゃ

1.53

↵↵↵

1.52

 rollercoaster

1.52

1.49

Activations Density 0.019%