INDEX

Explanations

pros and cons

The neuron detects headings or labels that introduce lists of advantages and disadvantages (e.g. “Pros,” “Cons,” “Advantages,” “Disadvantages”).

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

in

-1.79

to

-1.66

 conforman

-1.27

 diverso

-1.20

 afectan

-1.19

 indicate

-1.17

you

-1.16

 provides

-1.16

ticiones

-1.13

abbix

-1.13

POSITIVE LOGITS

and

1.85

 lavable

1.27

 propriétaires

1.24

 calendriers

1.23

 plastiques

1.20

Atlético

1.19

 élimin

1.19

 coûts

1.17

 étan

1.17

 réfrig

1.14

Activations Density 0.021%