INDEX

Explanations

The neuron fires on words that introduce or label propositions or viewpoints—e.g. “view,” “opinion,” “hypothesis,” and “premise.”

hypothesis and conclusion

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

Gön

-1.16

tasche

-1.12

bilden

-1.04

suche

-0.99

 такого

-0.99

เกง

-0.99

einer

-0.98

vég

-0.94

 کشاور

-0.93

IONES

-0.92

POSITIVE LOGITS

 that

3.70

 rằng

2.19

 bahwa

1.85

łbym

1.45

 about

1.33

認為

1.33

认为

1.29

 belief

1.27

 ότι

1.27

ว่า

1.24

Activations Density 0.046%