INDEX

Explanations

chances are that

The neuron detects expressions of likelihood or probability (e.g. “chances are,” “odds were against”).

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

⢅

-1.78

 seleccionados

-1.59

 北市

-1.44

~~~~~~

-1.42

 nød

-1.40

相当于

-1.37

簠

-1.37

 ブルゾン

-1.37

 $\&$

-1.35

哪儿

-1.33

POSITIVE LOGITS

you

1.63

،

1.56

みてください

1.52

—

1.52

𓆟

1.51

lawayo

1.51

he

1.51

acterísticas

1.49

謦

1.46

 premiere

1.45

Activations Density 0.007%