INDEX

Explanations

This neuron detects formal document section or heading phrases that end with the word “of,” such as “Summary of,” “Overview of,” or “Discussion of.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

its

-2.02

その

-1.48

把它

-1.30

手順

-1.30

それが

-1.28

irs

-1.27

could

-1.24

 solchen

-1.24

再一次

-1.23

が多く

-1.22

POSITIVE LOGITS

 each

1.48

 different

1.34

 other

1.32

 番目

1.30

 otros

1.27

主な

1.23

</h1>

1.19

</i>

1.17

 Each

1.14

 from

1.13

Activations Density 0.201%