INDEX

Explanations

connection is still

The neuron strongly activates on first-person self‐references (e.g. “I,” “my,” “me”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

وذ

0.64

(\{

0.51

كذا

0.51

ంద

0.50

(|

0.50

ည့်

0.50

 Unlike

0.49

Soon

0.49

uchsia

0.49

டுத்து

0.49

POSITIVE LOGITS

 still

5.31

 STILL

4.91

still

4.81

 Still

4.72

Still

4.66

仍然

4.38

 masih

4.34

仍

4.31

 vẫn

4.22

依然

4.19

Activations Density 1.769%