INDEX

Explanations

entity tags and code structure

The neuron fires on numeric tokens—especially measurements, counts, or other quantity values (e.g. dimensions, prices, pixel counts).

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

1.99

1.91

1.84

<0x0D>

1.72

<unused61>

1.67

1.66

<unused60>

1.65

1.64

↵

1.63

<unused62>

1.62

POSITIVE LOGITS

 अल्को

1.53

ेंद्रीय

1.47

ॕ

1.37

 fomentar

1.34

श्मीर

1.34

ल्ल्या

1.34

नावश्यक

1.34

 रक्षाबंधनाच्या

1.32

geoning

1.28

्हान

1.27

Activations Density 0.161%