Instruction Following on IFEval (Win Rate)
81.2Win RateRPO
Evaluation Results
| Method | Links | |
|---|---|---|
| RPOModel=Qwen2.5-7B, Baseline=SFT, LLM Judge=GPT-4o2025.06 | 81.2 | |
| RPOModel=LLaMA-3-8B, Baseline=SFT, LLM Judge=GPT-4o2025.06 | 80.5 | |
| RPOModel=Qwen2.5-7B, Baseline=SFT, LLM Judge=Claude-3.52025.06 | 80.4 | |
| RPOModel=LLaMA-3-8B, Baseline=SFT, LLM Judge=Claude-3.52025.06 | 79.6 | |
| RPOModel=Mistral-7B, Baseline=SFT, LLM Judge=GPT-4o2025.06 | 79.4 | |
| RPOModel=Mistral-7B, Baseline=SFT, LLM Judge=Claude-3.52025.06 | 78.3 | |
| RPOModel=FlanT5-XL, Baseline=SFT, LLM Judge=GPT-4o2025.06 | 73 | |
| RPOModel=FlanT5-XL, Baseline=SFT, LLM Judge=Claude-3.52025.06 | 71.8 | |
| RPOModel=FlanT5-Large, Baseline=SFT, LLM Judge=GPT-4o2025.06 | 68.9 | |
| RPOModel=FlanT5-Large, Baseline=SFT, LLM Judge=Claude-3.52025.06 | 67.5 | |
| RPOModel=Qwen2.5-7B, Baseline=KTO, LLM Judge=GPT-4o2025.06 | 63.9 | |
| RPOModel=FlanT5-Small, Baseline=SFT, LLM Judge=GPT-4o2025.06 | 63.5 | |
| RPOModel=LLaMA-3-8B, Baseline=KTO, LLM Judge=GPT-4o2025.06 | 63.1 | |
| RPOModel=Qwen2.5-7B, Baseline=KTO, LLM Judge=Claude-3.52025.06 | 63.1 | |
| RPOModel=FlanT5-Small, Baseline=SFT, LLM Judge=Claude-3.52025.06 | 62.8 | |
| RPOModel=LLaMA-3-8B, Baseline=KTO, LLM Judge=Claude-3.52025.06 | 62.5 | |
| RPOModel=Mistral-7B, Baseline=KTO, LLM Judge=GPT-4o2025.06 | 62.1 | |
| RPOModel=Mistral-7B, Baseline=KTO, LLM Judge=Claude-3.52025.06 | 61.3 | |
| RPOModel=FlanT5-XL, Baseline=KTO, LLM Judge=GPT-4o2025.06 | 61 | |
| RPOModel=FlanT5-XL, Baseline=KTO, LLM Judge=Claude-3.52025.06 | 60.2 | |
| RPOModel=Qwen2.5-7B, Baseline=DRO, LLM Judge=GPT-4o2025.06 | 59.5 | |
| RPOModel=FlanT5-Large, Baseline=KTO, LLM Judge=GPT-4o2025.06 | 59.3 | |
| RPOModel=LLaMA-3-8B, Baseline=DRO, LLM Judge=GPT-4o2025.06 | 59 | |
| RPOModel=Qwen2.5-7B, Baseline=DRO, LLM Judge=Claude-3.52025.06 | 58.7 | |
| RPOModel=Mistral-7B, Baseline=DRO, LLM Judge=GPT-4o2025.06 | 58.6 | |
| RPOModel=LLaMA-3-8B, Baseline=DRO, LLM Judge=Claude-3.52025.06 | 58.2 | |
| RPOModel=FlanT5-Large, Baseline=KTO, LLM Judge=Claude-3.52025.06 | 58.1 | |
| RPOModel=Mistral-7B, Baseline=DRO, LLM Judge=Claude-3.52025.06 | 57.8 | |
| RPOModel=FlanT5-XL, Baseline=DRO, LLM Judge=GPT-4o2025.06 | 57.2 | |
| RPOModel=FlanT5-Small, Baseline=KTO, LLM Judge=GPT-4o2025.06 | 57.1 | |
| RPOModel=FlanT5-Small, Baseline=KTO, LLM Judge=Claude-3.52025.06 | 56.4 | |
| RPOModel=FlanT5-XL, Baseline=DRO, LLM Judge=Claude-3.52025.06 | 56.4 | |
| RPOModel=FlanT5-Large, Baseline=DRO, LLM Judge=GPT-4o2025.06 | 55.8 | |
| RPOModel=FlanT5-Large, Baseline=DRO, LLM Judge=Claude-3.52025.06 | 54.6 | |
| RPOModel=FlanT5-Small, Baseline=DRO, LLM Judge=GPT-4o2025.06 | 54 | |
| RPOModel=FlanT5-Small, Baseline=DRO, LLM Judge=Claude-3.52025.06 | 53.2 |