INDEX

Explanations

now sees

The neuron strongly activates on first‐ and second‐person pronouns (e.g. “I,” “you,” “they”) in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 walnuts

0.55

ख्त

0.55

дзі

0.55

发出

0.54

ায়ের

0.52

 walnut

0.51

推進

0.51

 vykon

0.51

äden

0.51

 offers

0.50

POSITIVE LOGITS

 interacted

1.44

 interacts

1.43

 interacting

1.42

 interact

1.31

 만나

1.27

interacting

1.27

 interag

1.25

 ملاقات

1.24

 encounter

1.23

 interactions

1.21

Activations Density 0.602%