INDEX

Explanations

positive/negative/type semiconductors

The neuron fires strongly on technical numeric expressions—particularly decimal measurements, ranges, and parameter values (e.g. “150‐68,” “121.5‐34,” “p-22.25,” etc.).

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

-1.80

-1.62

-1.61

-1.57

you

-1.56

-1.52

...

-1.43

();

-1.36

</strong>

-1.34

 such

-1.30

POSITIVE LOGITS

 bañera

1.47

 เปิด

1.44

 reciclaje

1.38

𐐪

1.38

 tracción

1.35

 magasiner

1.34

𝘐

1.33

 universelle

1.32

 WITH

1.30

 delito

1.30

Activations Density 0.050%