INDEX

Explanations

code block endings

The neuron detects section or heading markers—tokens that introduce new labeled parts of a document (e.g. numbered questions, “Directions:”, “batches:”, “Q:”).

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

 पद्धतीने

0.84

 थांब

0.69

 प्रकारे

0.67

खाव

0.66

नातील

0.66

वतात

0.66

 सप्टेंबर

0.66

 लेणी

0.64

 मिळाल

0.64

णातील

0.63

POSITIVE LOGITS

1.12

↵

1.06

};

0.98

}//

0.92

↵↵

0.90

//

0.89

↵↵↵

0.89

;}

0.87

},

0.86

Activations Density 0.000%