INDEX

Explanations

proportional to, 权利, and equations

The neuron activates on reference and citation markers (e.g. bracketed numbers, journal codes, and period‐delimited author initials) in bibliographic entries.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 Postgres

0.95

 basicamente

0.93

 comida

0.92

 funcionalidades

0.91

 basiert

0.88

 prácticamente

0.88

 بسيط

0.88

 ايجاد

0.86

 خلينا

0.85

 nadie

0.84

POSITIVE LOGITS

①

0.89

﹣

0.79

②

0.79

ani

0.75

Ⅰ

0.73

ab

0.72

；

0.72

①

0.70

．

0.70

Activations Density 0.004%