INDEX

Explanations

series of examples

This neuron detects “series of” constructions, activating on the plural noun (e.g. articles, interviews, diseases, calls, events, lectures) that follows “series of.”

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

two

-1.41

 time

-1.33

 beginnings

-1.28

 präsenti

-1.27

ACIONAL

-1.25

 muchísimo

-1.25

了一个

-1.23

 often

-1.23

 this

-1.23

POSITIVE LOGITS

 других

1.57

⮑

1.51

 которые

1.47

 instituições

1.34

 olika

1.33

presumably

1.31

très

1.30

 that

1.30

 giúp

1.25

 kontaktu

1.23

Activations Density 0.097%