INDEX

Explanations

adjusting for factors

The neuron detects statistical control or adjustment phrases in research reporting (e.g. “controlling for,” “adjusted for,” “after controlling for”).

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

蛆

-1.49

 них

-1.48

 caucho

-1.45

 arcilla

-1.43

 málo

-1.41

 expliquer

-1.40

🏅

-1.39

varchar

-1.39

 réessayer

-1.36

 collectible

-1.35

POSITIVE LOGITS

for

2.11

欵

1.71

↵

1.69

 also

1.59

 that

1.49

 respecting

1.43

et

1.41

}{

1.41

 elaborated

1.40

 رسد

1.37

Activations Density 0.069%