INDEX

Explanations

thanks for support or actions

This neuron fires on tokens in expressions of thanks or acknowledgments—especially the preposition “for” in “thank you for…”–style phrases.

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

by

-2.17

 Those

-2.13

 people

-1.98

one

-1.98

 sophisticated

-1.97

 When

-1.95

 multifaceted

-1.92

 only

-1.91

５

-1.90

年も

-1.89

POSITIVE LOGITS

 praktische

2.11

providing

1.95

 evitando

1.95

 causando

1.93

隂

1.88

 entrando

1.80

loggen

1.80

 suscit

1.80

 horloge

1.79

tting

1.78

Activations Density 0.016%