INDEX

Explanations

starting position

This neuron detects procedural step indicators—words that mark stages or sequence in instructions (e.g., “first,” “initially,” “then,” “to …,” “anterograde,” “retrograde”).

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

 נוסף

-0.91

PENUTUP

-0.91

 updateUser

-0.89

 felpa

-0.87

Dentro

-0.87

맙

-0.84

additional

-0.84

Soph

-0.82

য

-0.82

interno

-0.82

POSITIVE LOGITS

 first

2.88

 starting

2.17

先

2.09

 сначала

2.03

 eerst

1.97

 ابتدا

1.96

 start

1.95

 initially

1.90

เริ่ม

1.87

 zunächst

1.85

Activations Density 0.082%