INDEX

Explanations

relative clauses starting with which

The neuron fires on “all of … which … are …” style quantifier phrases that specify a total count and a subset count or percentage.

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

镲

-1.87



-1.54

......”

-1.48

 akhirnya

-1.42

 even

-1.41

みたら

-1.41

trooper

-1.39

”

-1.38

POSITIVE LOGITS

 děti

1.78

焄

1.77

bingen

1.68

 trygg

1.66

 recaud

1.62

 rät

1.59

🢒

1.56

 mergul

1.52

伈

1.49

 MÁS

1.48

Activations Density 0.010%