INDEX

Explanations

asking questions

The neuron detects mentions of asking or answering questions and Q&A session cues (e.g., “asking,” “questions,” “ask,” “answer”).

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

 deacetyl

0.56

 linings

0.54

 चढ़ा

0.51

 halides

0.50

 coatings

0.48

បរ

0.48

 condam

0.48

 चढ़

0.47

覀

0.47

 studs

0.47

POSITIVE LOGITS

 questions

3.95

questions

3.78

 Questions

3.59

 질문

3.56

Questions

3.48

 Fragen

3.48

 QUESTIONS

3.38

 pertanyaan

3.34

 preguntas

3.31

 вопросы

3.27

Activations Density 1.058%