INDEX

Explanations

introduction to specific concepts

This neuron fires on occurrences of the token “labels,” especially when it appears as a field or keyword in code or data definitions.

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

🤺

-1.57

吮

-1.35

Previews

-1.29

 their

-1.26

 preciosas

-1.26

程度上

-1.25

ثيق

-1.22

 trataba

-1.22

 maison

-1.20

engkapnya

-1.20

POSITIVE LOGITS

ными

1.51

 were

1.38

 confezione

1.34

>`;

1.33

ющих

1.33

1.30

Πηγή

1.29

是

1.26

確認

1.26

 thing

1.23

Activations Density 0.000%