INDEX

Explanations

quotation mark followed by a word

The neuron fires strongly at the start of direct‐speech reply lines (i.e. the opening tokens—quotes and speaker pronouns—of interview or dialogue responses).

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

 their

-1.95

new

-1.95

-1.85

and

-1.66

 both

-1.63

an

-1.41

any

-1.41

 unique

-1.38

 präsenti

-1.37

 local

-1.30

POSITIVE LOGITS

we

1.95

轭

1.55

 vectorielles

1.51

骉

1.50

因为它

1.47

ABETH

1.42

 appelée

1.42

 haci

1.41

 bronzo

1.40

̓

1.40

Activations Density 0.003%