INDEX

Explanations

hotspot or hot spots

The neuron consistently activates on occurrences of “hot spot(s)” (including hyphenated and concatenated forms).

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

HEIM

-0.72

چه

-0.72

אור

-0.70

 Tham

-0.69

prehensive

-0.68

 बज

-0.68

観

-0.67

Caixa

-0.66

 titulada

-0.66

 Litecoin

-0.65

POSITIVE LOGITS

 hotspots

5.00

 hotspot

4.91

hot

4.53

hot

4.22

Hot

3.92

Hot

3.86

 Hotspot

3.73

HOT

3.00

热点

2.98

HOT

2.95

Activations Density 0.103%