INDEX

Explanations

TV, HDMI, video outputs

The neuron strongly activates on numeric specification tokens—i.e. numbers (often alongside units or dimensions) used in technical or product‐spec contexts.

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

.”

-1.41

)}$.

-1.35

</sup>

-1.35

 даже

-1.34

 después

-1.32

!”

-1.26

さんは

-1.20

)$.

-1.20

».

-1.18

<h2>

-1.17

POSITIVE LOGITS

 största

1.13

 gemens

1.10

 bezit

1.07

hanced

1.06

 VIDEOS

1.05

 Cerve

1.03

 each

1.02

 SEQUENCE

1.02

why

1.01

ragende

1.00

Activations Density 0.050%