Inference-Aware Fine-Tuning for Best-of-N Sampling in Large Language Models(arxiv.org)
69 points by mfiguiere 3 days ago | discuss