INDEX

Explanations

method, technique, process

The neuron detects discourse markers for describing new experimental or procedural techniques—words like “method,” “process,” or “technique.”

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

mô

-1.54

 skóra

-1.41

škoda

-1.29

 bílé

-1.29

ıyordu

-1.23

 srdce

-1.23

spectra

-1.23

skrä

-1.21

 Charakter

-1.21

guigu

-1.20

POSITIVE LOGITS

1.59

或

1.50

}{

1.48

1.42

1.40

ā

1.37

or

1.34

ć

1.32

still

1.30

Activations Density 0.222%