INDEX

Explanations

when introduces statements of being or past states

The neuron is primarily detecting clause-boundary function words—common conjunctions and subordinators (e.g. “when,” “and,” “but,” “why,” “are”) that introduce or link clauses.

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

我把

-1.05

 jeho

-1.00

她说

-0.97

他说

-0.91

我说

-0.86

 seinen

-0.82

 ее

-0.82

 его

-0.82

它的

-0.79

 whose

-0.79

POSITIVE LOGITS

自己是

1.36

我们是

1.22

 siamo

1.22

我是

1.17

 somos

1.15

 jestem

1.13

Iam

1.11

 estamos

1.11

他是

1.10

 jesteś

1.07

Activations Density 0.116%