INDEX

Explanations

recovering from

The neuron activates on words and phrases describing recovery or recuperation—especially forms of “recover” (e.g. “recovering,” “recovery,” “recover from”).

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

</strong>

-2.23

</i>

-1.98

</h1>

-1.84

-1.64

 Prüfung

-1.63

-1.60

 January

-1.59

ANY

-1.57

 influencing

-1.56

POSITIVE LOGITS

⚣

2.06

鹩

2.06

’

1.80

峮

1.78

 аниме

1.73

 meilleure

1.70

BEAR

1.69

 malade

1.69

 russes

1.67

︲

1.67

Activations Density 0.019%