INDEX

Explanations

state or status

The neuron fires on mixed-case identifier tokens (e.g. CamelCase or PascalCase method, variable, or flag names).

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

of

-2.50

are

-1.38

 newfound

-1.31

ledad

-1.30

».

-1.20

 после

-1.20

щего

-1.20

«

-1.18

 nėra

-1.18

POSITIVE LOGITS

龶

1.42

in

1.42

asties

1.37

個性

1.33

 эксперимента

1.30

 ersche

1.30

 ohnehin

1.30

Even

1.28

 її

1.27

ilingual

1.27

Activations Density 0.028%