INDEX

Explanations

The neuron is recognizing list items and enumerations

This neuron strongly fires at the start of long, structured informational passages (e.g. multi‐step instructions or formal lists), effectively marking the beginnings of detailed, instructional text blocks.

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

́s

0.78

_{

0.70

 repay

0.69

 เพียง

0.69

 Sở

0.69

́t

0.68

ϲ

0.68

 หมาย

0.66

 ใด

0.66

vyd

0.64

POSITIVE LOGITS

ESOME

0.84

BoardManager

0.82

гү

0.80

好处

0.79

ματα

0.79

 येतो

0.78

ubine

0.78

 mines

0.77

 نہایت

0.76

塩素

0.76

Activations Density 0.000%