INDEX

Explanations

e followed by numbers or "Item"

The neuron is a detector for the single-letter token “e” (in both lowercase and uppercase).

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

}$

-2.69

és

-1.98

();

-1.98

衚

-1.94

 kommunizieren

-1.91

ו

-1.91

िक

-1.85

⬘

-1.84

)$

-1.80

觥

-1.79

POSITIVE LOGITS



1.99

Ⴈ

1.83

Preparazione

1.78

菝

1.76

 رغم

1.69

 และ

1.67

1.65

 simplification

1.65

1.63

 THEN

1.62

Activations Density 0.009%