INDEX

Explanations

personal attacks and insults

This neuron fires on direct personal-address pronouns and statements—especially “I” and “you”—marking utterances that address the speaker or listener.

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

to

-2.03

 tikra

-1.59

]}$

-1.56

ması

-1.52

 melaksanakan

-1.48

一阵

-1.44

 dilaksanakan

-1.44

$};

-1.43

 during

-1.41

atilde

-1.41

POSITIVE LOGITS

 that

2.63

you

2.19

 your

2.13

 this

1.73

 these

1.67

who

1.66

my

1.54

這個

1.52

 industriale

1.45

and

1.43

Activations Density 0.059%