INDEX

Explanations

mathematical expressions and calculations

The main thing this neuron does is detect numeric tokens (quantitative 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

arıyla

0.70

kében

0.68

ﮢ

0.66

ujukan

0.65

業務

0.63

研修

0.63

 профилакти

0.62

riss

0.62

तांत्रिक

0.61



0.61

POSITIVE LOGITS

0.90

 eight

0.86

0.85

 cancels

0.78

 four

0.78

 cancel

0.77

 seven

0.75

 numerator

0.74

 five

0.74

 subtract

0.73

Activations Density 0.022%