INDEX

Explanations

The neuron is selectively activating on tokens that are long multi‐digit decimal numbers.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

年の

0.71

ilerin

0.70

ಳಿ

0.68

сія

0.66

śnie

0.65

萃

0.64

fers

0.63

Pepper

0.63

主義

0.63

cud

0.63

POSITIVE LOGITS

ک

0.85

ج

0.78

ا

0.75

 immobilier

0.72

یک

0.71

গ

0.71

 שהוא

0.70

 положение

0.70

 боле

0.70

 longitudine

0.69

Activations Density 0.000%