INDEX

Explanations

describing a static condition

The neuron consistently activates on occurrences of the word “static” (especially as an adjective labeling methods, forces, storage, images, networks, etc.).

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

沵

-2.41

了

-1.95

歺

-1.91

 Denken

-1.85

 thing

-1.84

睃

-1.83

 torner

-1.77

 Bestimm

-1.75

糁

-1.75

le

-1.73

POSITIVE LOGITS

if

2.23

2.06

』(

1.76

（※

1.76

 semangat

1.75

我了

1.71

 pertumbuhan

1.63

 diligent

1.59

了解更多

1.59

any

1.59

Activations Density 0.012%