INDEX

Explanations

past, predecessors, and history

The neuron fires on words naming roles, groups, or historical contexts—e.g. “presidents,” “players,” “predecessor,” “history,” “throughout,” “before”—as they appear in list- or record-style descriptions.

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

Manfaat

-1.34

-1.31

or

-1.30

new

-1.27

شكرا

-1.26

 esas

-1.25

 hermoso

-1.22

şiv

-1.20

heds

-1.20

نتیجه

-1.18

POSITIVE LOGITS

who

2.13

 whom

1.84

of

1.45

 generations

1.35

=[];

1.28

 predecessors

1.28

UCIÓN

1.27

='';

1.22

 since

1.21

)(-

1.21

Activations Density 0.077%