INDEX

Explanations

replicate, mimic, emulate

The neuron detects words expressing imitation or reproduction (e.g., “replicate,” “replicates,” “mimicked,” “emulating”).

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

urazione

-1.84

[])

-1.67

;}

-1.55

薷

-1.52

//}

-1.50

//{

-1.50

 نفسها

-1.48

渼

-1.46

)?;

-1.45

 nuova

-1.44

POSITIVE LOGITS

1.77

 that

1.50

 what

1.44

{//

1.40

//

1.37

 withstand

1.34

 réglementation

1.34

iskt

1.34

 imitate

1.33

the

1.32

Activations Density 0.070%