INDEX

Explanations

linking something to something

This neuron detects relational-linking constructions where one thing is connected, traced, or matched to another (e.g. “traced back to,” “match … to,” “link … to”).

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

ers

-1.31

</h5>

-1.26

 این

-1.25

 nhưng

-1.23

 deze

-1.16

 куда

-1.16

__*/

-1.15

 заказа

-1.13

ца

-1.12

tifications

-1.11

POSITIVE LOGITS

 consentimiento

1.62

pantalón

1.48

 aseguró

1.47

attente

1.45

 Denna

1.42

الدرا

1.41

 연결

1.37

 detalló

1.37

 крайне

1.37

 fenómeno

1.37

Activations Density 0.041%