INDEX

Explanations

The neuron fires on dense technical or mathematical terminology—especially words like “equations,” “model,” “analysis,” “nonlinear,” and related scientific jargon.

scientific equations

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

aptation

-1.16

嵛

-1.15

 улуч

-1.13

 einiger

-1.13

ästi

-1.12

 verschiedener

-1.10

 colorir

-1.09

 jedes

-1.08

pecabe

-1.07

 kilka

-1.05

POSITIVE LOGITS

now

1.62

for

1.47

 equations

1.34

 כולל

1.18

༎

1.16

in

1.15

‵

1.15

⠐

1.10

fæ

1.10

 both

1.10

Activations Density 0.039%