INDEX

Explanations

subsequent actions or states

The neuron activates on subordinating words that introduce relative or adverbial clauses (e.g. who, after, while).

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

any

-1.88

 then

-1.68

all

-1.64

 will

-1.63

 what

-1.48

 EVERY

-1.42

 HUGE

-1.42

or

-1.42

 your

-1.37

 that

-1.36

POSITIVE LOGITS

 recém

1.43

 anfitrión

1.36

 alcune

1.34

 plusieurs

1.29

 recentemente

1.27

思います

1.26

"$_

1.26

 recently

1.25

 várias

1.24

⩕

1.24

Activations Density 0.041%