Rejection Sampling

الترشيح بالرفض

أسلوب لتوليد بيانات تدريب عالية الجودة بتوليد عدة مخرجات من النموذج والاحتفاظ فقط بالصحيحة حسنة التنسيق ورفض الباقي.

A method for generating high-quality training data by producing multiple model outputs and keeping only the correct, well-formatted ones while discarding the rest.

Also translated asأخذ العينات بالرفض، الاصطفاء بالرفض

First appears in this corpus in: Learning to Summarize from Human Feedback (2020)

Appears in these papers

DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning2025 in the sky ✦
Learning to Summarize from Human Feedback2020 in the sky ✦
Llama 2: Open Foundation and Fine-Tuned Chat Models2023 in the sky ✦
Llama 2: Open Foundation and Fine-Tuned Chat Models2023 in the sky ✦
The Llama 3 Herd of Models2024 in the sky ✦
The Llama 3 Herd of Models2024 in the sky ✦
Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned2022 in the sky ✦
Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned2022 in the sky ✦
Fast Inference from Transformers via Speculative Decoding2023 in the sky ✦
WebGPT: Browser-Assisted Question-Answering with Human Feedback2021 in the sky ✦
WebGPT: Browser-Assisted Question-Answering with Human Feedback2021 in the sky ✦