INDEX

Explanations

conjunctions introducing subordinate clauses

The neuron is triggered by numeric tokens, especially floating‐point numbers.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 though

0.84

可以说

0.77

Though

0.70

可以说是

0.68

 Though

0.66

 although

0.62

 reportedly

0.61

 employed

0.59

 graceful

0.57

 своему

0.57

POSITIVE LOGITS

that

1.69

 وأن

1.46

 that

1.45

 että

1.29

 rằng

1.27

 bahawa

1.22

że

1.20

 bahwa

1.20

 दैट

1.19

THAT

1.19

Activations Density 0.196%