INDEX

Explanations

assembly:, Diagnosis of, Painter Daniel, Q:

This neuron activates on the token “view,” flagging occurrences of the word “view” (e.g. in “viewData”).

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

espół

-1.70

 camion

-1.64

compañ

-1.53

 唯美

-1.47

 chauffage

-1.45

 galeria

-1.45

 ambassade

-1.45

 ако

-1.44

öffnung

-1.41

 disques

-1.38

POSITIVE LOGITS

but

1.70

1.52

 with

1.31

In

1.29

1.23

<bos>

1.23

so

1.21

1.20

but

1.19

 не

1.19

Activations Density 0.000%