INDEX

Explanations

each followed by a noun

The neuron strongly activates on the word “each,” i.e. occurrences of that quantifying determiner.

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

still

-1.49

and

-1.45

-1.38

</strong>

-1.38

 while

-1.34

 toujours

-1.31

in

-1.30

 still

-1.30

нено

-1.30

 februari

-1.29

POSITIVE LOGITS

 each

2.00

 каждом

1.68

 poszczegól

1.62

Each

1.59

 bawat

1.56

畢竟

1.41

各有

1.40

Zubereitung

1.37

بیش

1.35

INTRODUCCIÓN

1.31

Activations Density 0.071%