INDEX

Explanations

overshot wheel, undercut structures, Aldershot Town

The neuron selectively activates on compound words formed with the directional prefixes “over-” or “under-” (e.g. overshoot, oversupply, undercut, underclass, overreach).

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

へ行

-2.13

 Lorsque

-2.05

</b>

-1.91

 nouvel

-1.90

黉

-1.86

氿

-1.78

⚋

-1.77

 של

-1.76

だいたい

-1.73

hending

-1.72

POSITIVE LOGITS

↵

2.55

妡

2.39

᳝

2.28

ᜀ

2.16

2.09

鬟

2.02

𞤢

2.02

 antworte

2.02

 thoughtfully

1.95

雰

1.95

Activations Density 0.004%