TIGER-AI-Lab/RewardHarness

Self-evolving agentic reward framework for image-editing evaluation — 47.4% on EditReward-Bench from only 100 preference demos, no reward-model training. arXiv 2605.08703.

语言：Python · GitHub 星标：98 · Fork：9

话题：agentic · ai-agents · benchmark · evaluation · gemini · image-editing · llm-evaluation · llm-judge

在 GitHub 上查看原仓库