INDEX

Explanations

describing technical projects or problems

The neuron fires on Spanish (and closely related Portuguese) content-section labels—words like “pregunta,” “contenido,” “problema,” “código,” “ejemplo,” etc.—that mark parts of a question post.

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

It

-1.12

üns

-1.03

 greatest

-1.03

年は

-1.02

the

-1.00

it

-1.00

一个

-1.00

 parceria

-0.99

best

-0.98

oczes

-0.98

POSITIVE LOGITS

 that

2.78

 whose

2.13

 which

2.06

 που

1.69

who

1.59

 الذي

1.58

 که

1.57

 который

1.46

 которые

1.39

whose

1.37

Activations Density 0.014%