INDEX

Explanations

don't have to

This neuron fires on “have to” obligation constructions—i.e. the infinitive marker “to” (and its following verb) when it expresses necessity (“have to do,” “have to worry,” etc.).

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

the

-1.91

㊗

-1.88

ⵌ

-1.79

 смеси

-1.73

 schnellen

-1.73

诓

-1.66

あぁ

-1.60

やっぱ

-1.60

たまたま

-1.60

こども

-1.59

POSITIVE LOGITS

也不能

1.74

↵

1.63

and

1.59

czegó

1.57

 Each

1.57

 Dezember

1.56

Ա

1.51

 They

1.48

 наша

1.46

служба

1.46

Activations Density 0.012%