INDEX

Explanations

language separators

The neuron strongly activates on numeric tokens—especially digits and numbers (including floating‐point values).

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

们

0.61

恕

0.60

రువు

0.58

 എന്നിവ

0.57

們

0.57

ános

0.55

 除了

0.55

↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵

0.55

த்த

0.55

 ஆகியவற்ற

0.55

POSITIVE LOGITS

、

1.27

、

1.20

﹑

1.05

”、“

1.03

」「

0.96

 maupun

0.90

%、

0.88

､

0.88

 illetve

0.87

､

0.86

Activations Density 0.011%