INDEX

Explanations

past actions or possessions

The neuron fires on occurrences of the auxiliary “have” (and “has”) introducing a perfect‐tense construction.

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

he

-1.65

 versucht

-1.47

年に

-1.46

末

-1.45

()){

-1.45

}\

-1.42

 моих

-1.41

꜄

-1.41

ributor

-1.41

Don

-1.39

POSITIVE LOGITS

to

1.46



1.37

 urme

1.34

并将

1.34

ついで

1.32



1.29

おすすめの

1.26

yacht

1.26

тные

1.24

鱻

1.23

Activations Density 0.100%