INDEX

Explanations

code transformation and string manipulation

The neuron detects Python string‐manipulation operations—calls like `.decode()`, `.encode()`, `.lower()`, or reverse‐slicing (`[::-1]`).

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

 имел

-1.53

 sebagainya

-1.42

惴

-1.38

 필요

-1.31

boretum

-1.29

roir

-1.28

 STILL

-1.28

 belleza

-1.28

 reloadData

-1.28

 gewoon

-1.25

POSITIVE LOGITS

1.77

𓆜

1.41

=’

1.39

“[

1.38

裒

1.38

assium

1.35

咦

1.34

 forskjellige

1.34

橚

1.34

 backen

1.34

Activations Density 0.002%