INDEX

Explanations

switching to from

The neuron selectively activates on occurrences of “switch” and its inflected forms (e.g. switch, switches, switching, switched).

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

<bos>

-1.56

これは

-1.41

 vermag

-1.30

行きたい

-1.29

场比赛

-1.28

强的

-1.27

 заявил

-1.26

亵

-1.23

ズニー

-1.23

讫

-1.21

POSITIVE LOGITS

to

2.84

 into

2.20

 from

1.71

the

1.66

 إلى

1.47

 från

1.43

おわりに

1.40

 به

1.37

 از

1.34

 kategor

1.28

Activations Density 0.048%