Partition Function
ثابت التقسيم
Z(x): ثابت تسوية يضمن أن مجموع الاحتمالات يساوي واحداً. في DPO، يُلغى تلقائياً عند المقارنات الزوجية فلا حاجة لحسابه.
Z(x): a normalization constant ensuring probabilities sum to one. In DPO, it cancels automatically in pairwise comparisons, so it never needs to be computed.
Also translated asدالة التجزئة، ثابت التطبيع
First appears in this corpus in: Equation of State Calculations by Fast Computing Machines (1953)
Appears in these papers
- A Learning Algorithm for Boltzmann Machines1985in the sky ✦
- Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data2001in the sky ✦
- Direct Preference Optimization: Your Language Model Is Secretly a Reward Model2023in the sky ✦
- Equation of State Calculations by Fast Computing Machines1953in the sky ✦
- node2vec: Scalable Feature Learning for Networks2016in the sky ✦
- Generative Modeling by Estimating Gradients of the Data Distribution2019in the sky ✦