INDEX

Explanations

action after `and`

The neuron flags numeric tokens—especially multi‐digit and decimal numbers—anytime a quantity (e.g. “486.864,” “2031.051,” “902.215”) appears in the text.

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

ಂಟು

0.64

につなが

0.63

َي

0.61

َد

0.61

 పాల్గొన్నారు

0.59

<unused2223>

0.58

engaruhi

0.57

aran

0.55

 Isso

0.55

 شائع

0.55

POSITIVE LOGITS

 whilst

2.02

 during

1.91

 although

1.91

 despite

1.91

 after

1.91

 while

1.81

 setelah

1.80

หลังจาก

1.80

 when

1.77

 після

1.77

Activations Density 0.560%