INDEX

Explanations

success status

The neuron fires on structured “success” status indicators (e.g. JSON/XML fields like “success” or “IsSuccess” and their true/false values).

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

 have

-1.48

ும்

-1.29

睨

-1.23

 ...,

-1.22

ろし

-1.19

 their

-1.18

すと

-1.16

wego

-1.16

matisch

-1.16

ново

-1.15

POSITIVE LOGITS

Regardless

1.37

المشار

1.30

類

1.30

兩

1.29

を目指す

1.29

 presidenta

1.28

Really

1.27

 success

1.27

 results

1.27

 conclusion

1.27

Activations Density 0.032%