INDEX

Explanations

which/that/who followed by a verb

The neuron detects relative‐clause markers (words like which, that, who, etc.) introducing subordinate clauses.

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

赑

-1.48

 also

-1.43

kiz

-1.42

 españ

-1.41

中身

-1.40

 vocales

-1.40

 ملک

-1.38

!«

-1.37

пили

-1.36

 protože

-1.34

POSITIVE LOGITS

Initially

1.55

ainya

1.45

這麼

1.41

 таком

1.37

You

1.35

乾淨

1.34

reggio

1.32

ないし

1.32

かったので

1.30

 μαζί

1.30

Activations Density 0.103%