INDEX

Explanations

spend time on their

The neuron detects phrases that quantify how much time (often with numbers or superlatives like “most” or percentages) someone or something spends doing an activity.

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

as

-1.75

 only

-1.70

 milyen

-1.62

 hogyan

-1.53

 such

-1.41

}$.

-1.39

.<

-1.38

had

-1.35

 didn

-1.33

POSITIVE LOGITS

on

1.55

 produire

1.51

拚

1.48

 nuovi

1.41

꯳

1.41

 gagné

1.38

 benutzt

1.37

 новых

1.36

 nuove

1.34

︕

1.33

Activations Density 0.042%