INDEX

Explanations

man, woman, men, women

This neuron fires on gendered human nouns (e.g., “man,” “woman,” “women”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

</h4>

-1.83



-1.82

 terv

-1.73

了许多

-1.71

 investissement

-1.70

</sup>

-1.66

 tesettür

-1.63

 kaos

-1.61

 abiye

-1.59

Nicht

-1.58

POSITIVE LOGITS

 They

2.09

</i>

1.86

 This

1.76

 called

1.70

1.66

 WITH

1.59

 said

1.59

<h2>

1.55

 told

1.54

TO

1.52

Activations Density 0.004%