INDEX

Explanations

shutdown or downtime

The neuron primarily detects numeric tokens (e.g. multi‐digit numbers, dates, counts, percentages).

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

 உருவாக்க

0.64

 McManus

0.59

 बतौर

0.57

 साझेदारी

0.57

ბ

0.57

 cins

0.56

 tray

0.55

 деб

0.55

 விய

0.55

cub

0.54

POSITIVE LOGITS

 shutdown

1.38

 Shutdown

1.29

 shutdowns

1.27

shutdown

1.25

Shutdown

1.23

 shut

1.17

 shuts

1.16

 تعط

1.13

休み

1.11

 ছুটি

1.04

Activations Density 0.287%