INDEX

Explanations

rather than, because, as, with, like

The neuron detects discourse‐linking words (subordinating conjunctions and adverbial connectives) such as “because,” “since,” “as,” “when,” “rather,” and “like.”

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

itados

-1.66

occurs

-1.59

 underlines

-1.54

 brisa

-1.54



-1.53

፨

-1.53

刄

-1.52

Preguntas

-1.51

 enviada

-1.50

 dénon

-1.50

POSITIVE LOGITS

for

2.08

1.84

to

1.71

De

1.62

 when

1.55

at

1.50

 with

1.48

–

1.42

as

1.42

an

1.39

Activations Density 0.241%