INDEX

Explanations

effort or lack thereof

The neuron fires on emotionally or rhetorically emphatic words and intensifiers—e.g. “thanks,” “damn,” “even,” “actually,” or verbs like “trying,” “express,” “give,” “bother,” “attempt.”

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

 consommateurs

-1.06

 pouvoirs

-1.05

 nuages

-1.04

 cœurs

-1.03

 milliers

-1.03

 dommages

-1.03

 centaines

-1.02

 renseignements

-1.01

 unele

-1.01

 dégâts

-0.99

POSITIVE LOGITS

to

1.66

or

1.47

 или

1.20

 porque

1.17

 Because

1.16

 because

1.10

any

1.09

ーめん

1.02

Because

1.01

 Since

0.99

Activations Density 0.040%