INDEX

Explanations

return statements

The neuron detects imperative control‐flow keywords that immediately exit or return from a function or program (e.g. “return,” “exit,” “quit,” etc.).

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

-1.00

-0.99

-0.98

 lorsque

-0.96

 الشرطة

-0.91

-0.88

-0.86

 لي

-0.85

phat

-0.84

 Geburt

-0.82

POSITIVE LOGITS

mismo

0.99

rów

0.97

}});

0.94

☞

0.94

 halte

0.91

𝓫

0.90

interrupted

0.90

𝔱

0.90

 this

0.89

🐕

0.88

Activations Density 0.001%