INDEX

Explanations

without a single

The neuron consistently activates on the word “single.”

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

</sub>

-1.57

 asamblea

-1.52

少し

-1.52

妡

-1.51

 également

-1.40

 Privacidade

-1.39

 Accessoires

-1.38

螄

-1.38

ENNIS

-1.36

趼

-1.35

POSITIVE LOGITS

 those

1.60

 отлично

1.52

 tandis

1.51

するのは

1.47

ゎ

1.44

都是在

1.42

 verschied

1.41

it

1.39

 gewiß

1.39

âtel

1.38

Activations Density 0.017%