INDEX

Explanations

I'm expressing personal states

The neuron fires on first‐person self‐references, i.e. tokens like “I,” “I’m,” and other “I”-based pronouns.

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

 宝宝

-1.18

様でした

-1.17

￭

-1.14

 itſelf

-1.13

игра

-1.11

 modifi

-1.09

诸葛

-1.08

んですよね

-1.07

興味

-1.06

 fenti

-1.06

POSITIVE LOGITS

</h2>

1.65

 will

1.54

 into

1.36

 czasie

1.34

1.30

</h1>

1.30

〉

1.30

OrNil

1.29

 ultimi

1.27

 обзор

1.27

Activations Density 0.049%