INDEX

Explanations

highs and lows

The neuron detects mentions of quantitative extrema—phrases indicating numeric highs, lows, records, or comparative measurements (e.g., “rose to,” “lowest since,” “all-time high”).

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

of

-1.81

 sourire

-1.66

 môžete

-1.44

 énergie

-1.42

 землю

-1.41

-1.40

 budete

-1.40

 chimiques

-1.39

 byste

-1.38

₂,

-1.38

POSITIVE LOGITS

 empezó

1.52

1.43

1.30

we

1.29

 their

1.27

 demanded

1.26

 había

1.25

⸫

1.24

 hardly

1.24

可以

1.23

Activations Density 0.031%