INDEX

Explanations

walls and electrodes

The neuron detects mentions of structural boundary terms—especially “walls,” “surfaces,” “plates,” and similar physical interface words—in technical or scientific descriptions.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 kablo

-1.47

 butik

-1.37

 adal

-1.31

 boya

-1.30

ledes

-1.30

 praktik

-1.28

趿

-1.28

 seksi

-1.27

 kazak

-1.27

她们

-1.27

POSITIVE LOGITS

 Generally

1.32

if

1.31

If

1.29

 reír

1.28

 oznacza

1.27

1.25

 said

1.24

 durch

1.24

at

1.23

 Often

1.23

Activations Density 0.071%