INDEX

Explanations

places and events

The neuron strongly activates on proper‐name tokens—titles of books and journals, brand or product names, place names, event names, etc.—i.e. named entities.

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

 your

-1.42

還會

-1.33

you

-1.24

ṓ

-1.17

ができます

-1.05

You

-1.04

or

-1.03

are

-1.02

すっ

-1.00

还会

-1.00

POSITIVE LOGITS

1.43

 وكانت

1.19

mà

1.19

 وكان

1.14

 excelente

1.13

 Había

1.13

 таки

1.12

 Algunas

1.11

 yesterday

1.09

othed

1.09

Activations Density 0.094%