Matched compute reverses dense-reward gains over GRPO
Across 1,668 runs, a Monte-Carlo process reward trailed GRPO by 14.3 points at equal tokens on Qwen2.5-1.5B.
1 stories, newest first · one card per story
Across 1,668 runs, a Monte-Carlo process reward trailed GRPO by 14.3 points at equal tokens on Qwen2.5-1.5B.
Today 1.5K readers · 2.4K views · 757 reads · all time 189K views