INDEX

Explanations

leaves a trace

This neuron fires on mentions of something being left, left behind, or remaining (e.g., “left,” “leaving,” “behind,” “by”).

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

 cinque

-1.74

橚

-1.62

-1.61

儛

-1.59

鳎

-1.56

፱

-1.54

却不

-1.53



-1.50

Ꮷ

-1.47

鸸

-1.47

POSITIVE LOGITS

שלום

1.45

 放題

1.40

 jūsų

1.37

RESUMEN

1.36

 Generally

1.35

Kui

1.34

婚礼

1.33

 також

1.32

 tenacious

1.31

 Probably

1.30

Activations Density 0.045%