INDEX

Explanations

consume or produce

The neuron detects occurrences of the root “consume” (in forms like consume, consumed, consumer, 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

！”

-1.79

 they

-1.58

?”

-1.54

!”

-1.53

…”

-1.51

how

-1.50

珧

-1.48

。”

-1.42

？”

-1.42

婁

-1.40

POSITIVE LOGITS

 from

1.67

 producent

1.65

 Эти

1.64

 That

1.63

 These

1.59

飧

1.58

 güne

1.56

1.52

We

1.49

koszulka

1.46

Activations Density 0.059%