INDEX

Explanations

at least / exactly

The neuron activates on numeric tokens—especially multi-digit numbers and decimals—highlighting occurrences of numerical values in the text.

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

 लगभग

1.05

 почти

1.02

 تقریباً

1.01

nearly

0.98

 nearly

0.97

almost

0.96

 prawie

0.95

 લગભગ

0.94

 hampir

0.91

幾乎

0.91

POSITIVE LOGITS

两种

0.82

 three

0.78

兩種

0.76

two

0.74

 drei

0.73

dua

0.73

两个

0.72

三种

0.72

 zwei

0.69

 trois

0.68

Activations Density 0.297%