INDEX

Explanations

item number or identifier

The neuron detects references to specific product “item” entries, especially the word “item” and its accompanying item‐ID numbers.

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

Then

-1.73

-1.55

 then

-1.55

wakilan

-1.50

):

-1.49

-1.45

하는

-1.43

."

-1.41

Also

-1.41

).

-1.40

POSITIVE LOGITS

 atletas

1.59

 Ordem

1.57

 pelea

1.57

孥

1.55

􏰀

1.53

汌

1.52

 tantos

1.52

itism

1.50

埽

1.50

 sillas

1.49

Activations Density 0.018%