INDEX

Explanations

past actions or states

The neuron is primarily triggered by past‐tense verbs and temporal framing words that mark personal experiences (e.g. “had,” “was,” “for,” “at,” “time”).

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

 otros

-1.57

Which

-1.43

 outros

-1.39

其他的

-1.37

youts

-1.36

 project

-1.33

łka

-1.32

endish

-1.31

ametro

-1.28

 возмо

-1.28

POSITIVE LOGITS

by

1.91

 with

1.81

 when

1.80

one

1.67

we

1.66

as

1.60

on

1.58

 during

1.55

 During

1.44

 while

1.43

Activations Density 0.010%