INDEX

Explanations

ok

The neuron is activated by numeric tokens—especially floating‐point numbers (e.g. “1855.6328”, “652.2424”)—so it’s detecting decimal number expressions.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

<start_of_image>

1.23

↵↵

1.17

 继续访问

0.62

</h2>

0.61

 (…)

0.58

</h3>

0.56

ख्या

0.55

 […]

0.54

</h4>

0.52

Jes

0.51

POSITIVE LOGITS

0.55

↵

0.55

천

0.52

 पाठ्यक्रम

0.50

дома

0.50

طلاق

0.50

0.49

 જિલ્

0.49

\,\

0.49

0.48

Activations Density 0.099%