INDEX

Explanations

object attribute access

The neuron fires on error messages or code snippets mentioning missing attributes or objects—especially “has no attribute”/“AttributeError” patterns.

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

 منم

-1.36

 عزیزم

-1.35

再一次

-1.19

 ਵਿੱਚ

-1.17

ritic

-1.15

 dichas

-1.13

 yaitu

-1.09

นะครับ

-1.09

 will

-1.08

 आपको

-1.08

POSITIVE LOGITS

これ

1.34

 what

1.29

山东

1.26

uttgart

1.23

閱讀

1.22

our

1.22

For

1.20

從

1.20

 Destroyer

1.16

Would

1.14

Activations Density 0.073%