INDEX

Explanations

Second followed by historical context

The neuron consistently detects the occurrence of the ordinal “Second” (e.g. in “Second World War,” “Second National Survey,” etc.).

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

“

-2.86

-2.28

-2.14

 この

-2.13

 vienen

-1.89

но

-1.88

-1.85

「

-1.83

his

-1.78

 been

-1.73

POSITIVE LOGITS

 Importantly

2.69

 Whereas

2.30

);

2.25

袿

2.23

 Obviously

2.23

 illustratie

2.17

 Therefore

2.16

 beslu

2.14

ayaquil

2.09

！！”

2.09

Activations Density 0.018%