INDEX

Explanations

start or continue

The neuron is detecting personal‐address usages—especially first‐ and second‐person pronouns (“I,” “me,” “you”) in questions or requests.

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

Истори

-1.59

 zatem

-1.58

 réve

-1.55

 到着

-1.49

 novamente

-1.49

 których

-1.49

nhà

-1.46

 conservé

-1.46

 desplaz

-1.46

甭

-1.45

POSITIVE LOGITS

in

2.42

it

2.16

are

2.11

1.95

is

1.63

 different

1.59

 have

1.57

alties

1.56

to

1.54

on

1.47

Activations Density 0.780%