INDEX

Explanations

for

The neuron fires on apostrophe characters used in contractions (e.g., “I’m,” “don’t,” “you’re”).

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

of

-2.17

会儿

-1.94

ଵ

-1.84

 после

-1.72

 surla

-1.69

⦊

-1.68

taya

-1.67

猹

-1.65

柘

-1.62

蓼

-1.55

POSITIVE LOGITS

 then

2.11

1.88

 also

1.73

 like

1.71

なども

1.70

曾經

1.70

 that

1.68

饜

1.66

 such

1.66

archiwizowane

1.59

Activations Density 0.060%