INDEX

Explanations

get; set;

The neuron detects imperative or mutation‐related keywords in code, such as “set,” “write,” and “OR REPLACE.”

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

ỏng

-1.31

𖤣

-1.31

 girar

-1.26

뭘

-1.21

หรือ

-1.20

 nadru

-1.20

 ludzie

-1.18

 golpea

-1.18

 gitara

-1.17

 muitos

-1.16

POSITIVE LOGITS

of

1.80

1.23

 спосо

1.14

there

1.10

now

1.09

та

1.06

1.05

for

1.02

1.01

 were

1.00

Activations Density 0.005%