INDEX

Explanations

programming keywords

The neuron fires on common code‐declaration keywords and modifiers (e.g. “internal,” “override,” “public,” “class,” “fun,” 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

を変更

-1.54

 tidligere

-1.34

সূত্র

-1.30

 medarbe

-1.29

 måneder

-1.26

前回の

-1.25

今年の

-1.23

 katika

-1.22

 måned

-1.20

 medlemmer

-1.20

POSITIVE LOGITS

 from

1.41

how

1.30

</strong>

1.30

</em>

1.27

 must

1.22

 не

1.22

</i>

1.20

未能

1.13

 optionally

1.13

と

1.10

Activations Density 0.063%