INDEX

Explanations

per year, Jewish lore, catastrophic event

This neuron responds to numeric information—measurements, quantities, statistics—and highlights number tokens (and occasionally numeric-related terms) 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

 Commencez

0.93

 иллюстра

0.86

 alteraciones

0.86

ете

0.85

 erreurs

0.85

 искусства

0.84

 самое

0.83

 venezolano

0.82

 étrangers

0.81

 conectados

0.81

POSITIVE LOGITS

od

0.76

pp

0.70

ua

0.70

ami

0.69

so

0.68

pt

0.68

 here

0.67

fontsize

0.67

hope

0.67

Activations Density 0.001%