INDEX

Explanations

possessive pronouns

The neuron selectively activates on third-person possessive markers (e.g., “his,” “her,” “their”), identifying words that denote ownership or belonging.

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

 onderdelen

-1.42

 verschillen

-1.34

its

-1.32

 dezelfde

-1.31

fordshire

-1.29

了什么

-1.26

 jutaan

-1.20

quelize

-1.18

 golongan

-1.17

by

-1.16

POSITIVE LOGITS

 latest

1.62

 heilige

1.27

 gefe

1.22

 másik

1.22

 heilig

1.21

 senaste

1.20

Whe

1.17

gebracht

1.16

});*/

1.16

######

1.16

Activations Density 0.061%