INDEX

Explanations

physical surfaces and water

The neuron strongly activates on concrete location or surface nouns (e.g. field, table, floor, kitchen), i.e. words naming physical settings.

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

-2.17

’

-2.13

-1.82

-1.81

If

-1.81

-1.77

’)

-1.76

卨

-1.74

that

-1.70

-1.65

POSITIVE LOGITS

of

2.77

for

1.99

quelize

1.81

 adım

1.76

 lunghe

1.71

 conjunction

1.70

 teşekkür

1.68

 Polícia

1.66

viembre

1.63

тельства

1.63

Activations Density 0.085%