INDEX

Explanations

semaphore flags / traffic lights

The neuron is detecting occurrences of the term “Semaphore” (i.e. references to semaphore constructs).

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

也在

-2.13

.”

-2.05

⢮

-2.03

 Einzelnen

-1.94

 sheer

-1.89

 extremely

-1.84

 swat

-1.82

 sorely

-1.78

 hatte

-1.76

ти

-1.75

POSITIVE LOGITS

}$

2.69

</th>

2.66

</sub>

2.59

);

2.42

雰

2.25

2.23

</h1>

2.19

徬

2.13

𖥧

2.08

锕

2.03

Activations Density 0.005%