INDEX

Explanations

Legal and code references

The neuron fires most strongly on comment‐block annotation tokens (e.g. leading “*” markers and “@” tags) in code documentation.

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

 розпові

-0.89

 отече

-0.85

ハロウィン

-0.83

 стаби

-0.81

 wyróż

-0.81

竟然是

-0.79

 optimisation

-0.77

วาง

-0.76

сро

-0.76

ntgen

-0.75

POSITIVE LOGITS

 októ

0.81

 mikrofon

0.81

 koje

0.80

 keram

0.76

وبة

0.76

aix

0.75

 procesu

0.74

>({

0.73

($("#

0.73

ürn

0.73

Activations Density 0.002%