INDEX

Explanations

mathematical notation and technical terms

The neuron activates on standalone numeric tokens (i.e. sequences of digits) wherever they appear in the text or math.

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

 assist

-0.85

 lagar

-0.75

 assumptions

-0.73

 acceptability

-0.73

 LEGEND

-0.73

 MARINE

-0.72

syk

-0.71

 Judson

-0.71

 notwithstanding

-0.71

 accept

-0.70

POSITIVE LOGITS

frac

0.98

sqrt

0.98

fr

0.76

avent

0.75

 Forex

0.74

fő

0.73

mbox

0.73

Big

0.73

 sqrt

0.72

 découvert

0.72

Activations Density 0.020%