INDEX

Explanations

code delimiters

The neuron is keyed to iteration constructs in code—especially “for” loops and related tokens (like “in”, collection names, and braces) indicating a loop or traversal.

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

↵↵

-1.46

);

-1.26

oquia

-1.20

 stava

-1.16

ятся

-1.15

erialized

-1.14

[],

-1.13

一旁

-1.10

一旁的

-1.09

 measures

-1.08

POSITIVE LOGITS

including

1.32

 thier

1.27

they

1.25

呦

1.23

 Soit

1.22

 their

1.20

pueden

1.18

越高

1.18

喽

1.17

哟

1.16

Activations Density 0.008%