INDEX

Explanations

raw followed by varied nouns

The neuron activates on occurrences of the adjective “raw.”

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

´

-1.98

 eſſ

-1.94

")

-1.84

-1.71

螓

-1.60

ted

-1.59

⟞

-1.55

￬

-1.54

フック

-1.50

POSITIVE LOGITS

raw

1.72

Although

1.70

 beweisen

1.69

вів

1.64

 температуре

1.62

日を

1.61

 behaupten

1.59

 meningkat

1.57

Depending

1.56

 menurun

1.55

Activations Density 0.012%