INDEX

Explanations

root development; avoid freezing

The neuron activates on occurrences of the string “root,” whether talking about actual plant roots, root vegetables, or the concept of rooting (e.g. “roots,” “rooted,” “root vegetables”).

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

 repaint

0.56

 blem

0.54

 broch

0.54

 cabins

0.52

 Crisp

0.52

শিষ্ট

0.52

ാര്യ

0.52

 checklists

0.52

cabin

0.51

 cabinets

0.50

POSITIVE LOGITS

 roots

2.31

roots

2.14

 Roots

2.14

 root

2.13

Root

2.09

Roots

2.09

根

2.06

root

2.03

 Root

2.02

ROOT

1.88

Activations Density 0.190%