INDEX

Explanations

focus or priority

The neuron responds to high-level discourse tokens—common clause- or sentence-boundary markers and intensifiers (e.g. “but,” “I,” “best,” “more,” “rather,” 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

%$$

-0.91

 почина

-0.84

 secretions

-0.82

Selections

-0.80

ार्थ

-0.79

 Tere

-0.78

𝘉

-0.77

 zarówno

-0.76

 sultan

-0.75

 citadel

-0.74

POSITIVE LOGITS

 focus

2.66

focus

2.39

Focus

2.08

 Focus

1.85

 focuses

1.79

 concentrate

1.70

 focusing

1.67

 focused

1.65

 fokus

1.61

FOCUS

1.46

Activations Density 0.043%