INDEX

Explanations

organization names and addresses

The neuron strongly activates on numeric tokens and numeric sequences (e.g. street numbers, section numbers, and other multidigit/decimal expressions).

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

 inconsist

0.29

faits

0.25

 salespeople

0.25

 giveaways

0.24

윙

0.24

 অশ্লীল

0.24

collectibles

0.24

 blatantly

0.24

क्रोश

0.24

 affidavits

0.24

POSITIVE LOGITS

 University

0.44

 Universität

0.41

 Universidad

0.38

 Institute

0.38

 Universidade

0.37

 Université

0.37

University

0.36

 Universit

0.36

 Laboratory

0.35

 Sciences

0.35

Activations Density 0.216%