INDEX

Explanations

make the most of

The neuron consistently fires on the word “most,” especially in “make the most of” constructions.

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

-1.95

睒

-1.92

</b>

-1.84

you

-1.81

 schein

-1.55

{$\

-1.53

を果た

-1.52

 woes

-1.52

-1.51

 ciudadana

-1.49

POSITIVE LOGITS

of

2.73

matitis

1.90

。“

1.83

")

1.81

 также

1.80

“

1.80

 expérience

1.72

 also

1.71

</h3>

1.70

取り付け

1.69

Activations Density 0.005%