INDEX

Explanations

everyday objects

The neuron fires on bits of “how‐to” or DIY instructional text—imperative verbs, quantities, measurements, and step indicators (e.g. Using, Just lay, 100%, 5‐gallon, etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

hülle

-0.94

Computers

-0.84

imų

-0.84

 preservatives

-0.82

 Biografie

-0.81

Thurs

-0.80

 Thurs

-0.79

ایا

-0.78

amel

-0.78

第一

-0.77

POSITIVE LOGITS

 household

1.57

 kitchen

1.39

old

1.31

 inexpensive

1.14

 antiguo

1.13

 cardboard

1.13

 supermarket

1.13

 coffee

1.12

 cheap

1.10

household

1.09

Activations Density 0.068%