INDEX

Explanations

repeated reading of "here"

This neuron detects generic link-anchor words (like “this” and “here”) that point to referenced resources or related posts.

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

一颗

-1.34

to

-1.34

ANCHE

-1.27

 fundido

-1.22

žne

-1.20

 Swine

-1.18

ávání

-1.17

年龄

-1.16

semaphore

-1.16

一旁

-1.15

POSITIVE LOGITS

}{

1.43

];

1.34

»،

1.30

alism

1.29

They

1.27

);

1.26

 просто

1.25

},

1.23

 которые

1.23

igth

1.22

Activations Density 0.025%