INDEX

Explanations

quality of items

The neuron activates on occurrences of the word “quality.”

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

餍

-1.72

 interieur

-1.59

誑

-1.57

ſy

-1.56

 eléctrico

-1.56

 tabak

-1.56



-1.54

 plafon

-1.52

 タンク

-1.52

豨

-1.51

POSITIVE LOGITS

1.72

<strong>

1.70

());

1.69

</em>

1.63

However

1.60

If

1.56

</

1.55

1.53

).

1.52

镌

1.52

Activations Density 0.020%