INDEX

Explanations

family members and roles

The neuron activates strongly on nouns that label or identify people (e.g. demographic or relational roles like “woman,” “teen,” “daughter,” “dad,” “soldier,” “friends”).

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

 from

-1.61

 such

-1.41

 összefoglaló

-1.38

for

-1.35

and

-1.32

 összefoglal

-1.32

postmedia

-1.16

 which

-1.14

 what

-1.11

 információ

-1.08

POSITIVE LOGITS

 methode

1.33

 genannten

1.16

 باسلام

1.13

 erwähnten

1.11

 observed

1.11

 intent

1.09

iddish

1.09

⢓

1.08

on

1.07

 gesagt

1.07

Activations Density 0.177%