INDEX

Explanations

names of people and institutions

The neuron fires on document‐structuring or formatting markers—long runs of punctuation (e.g. “~~~”), special symbols (e.g. bullets “â¢”), and capitalized section or code identifiers.

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

蓠

-2.91

苤

-2.89

在他的

-2.81

</h4>

-2.81

"#

-2.70

崀

-2.63

 alguns

-2.53

 YEAR

-2.47

 exista

-2.47

"$

-2.45

POSITIVE LOGITS

?”

2.78

罌

2.63

</i>

2.36

 longstanding

2.33

-“

2.30

 When

2.30

 nią

2.30

靄

2.30

For

2.22

 revital

2.22

Activations Density 0.034%