Instruction Following on IFEval (IF-RT, IF-FA, Avg. IF)
76.26IF-RT ScoreIF-RT chkpt.
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| IF-RT chkpt.Family=Qwen 3, Size (B)=82026.02 | 76.26 | 47.48 | 61.87 | |
| Staged DecodingFamily=Qwen 3, Size (B)=82026.02 | 76.26 | 86.15 | 81.21 | |
| IF-RT chkpt.Family=Qwen 3, Size (B)=142026.02 | 68.35 | 54.38 | 61.36 | |
| Staged DecodingFamily=Qwen 3, Size (B)=142026.02 | 68.35 | 86.99 | 77.67 | |
| IF-RT chkpt.Family=Qwen 3, Size (B)=42026.02 | 65.29 | 40.89 | 53.09 | |
| Staged DecodingFamily=Qwen 3, Size (B)=42026.02 | 65.29 | 79.02 | 72.15 | |
| IF-RT chkpt.Family=Qwen 3, Size (B)=1.72026.02 | 56.18 | 37.17 | 46.67 | |
| Staged DecodingFamily=Qwen 3, Size (B)=1.72026.02 | 56.12 | 63.67 | 59.89 | |
| IF-RT chkpt.Family=Phi 4, Size (B)=142026.02 | 46.64 | 48.56 | 47.6 | |
| Staged DecodingFamily=Phi 4, Size (B)=142026.02 | 46.64 | 65.59 | 56.12 | |
| IF-RT chkpt.Family=Phi 4, Size (B)=3.82026.02 | 41.67 | 37.95 | 39.81 | |
| Staged DecodingFamily=Phi 4, Size (B)=3.82026.02 | 41.67 | 54.2 | 47.93 | |
| BaselineFamily=Phi 4, Size (B)=142026.02 | 41.55 | 91.79 | 66.67 | |
| IF-FA chkpt.Family=Phi 4, Size (B)=142026.02 | 41.13 | 89.63 | 65.38 | |
| Avg. IF chkpt.Family=Phi 4, Size (B)=142026.02 | 40.95 | 91.01 | 65.98 | |
| IF-FA chkpt.Family=Phi 4, Size (B)=3.82026.02 | 39.93 | 56.59 | 48.26 | |
| Avg. IF chkpt.*Family=Phi 4, Size (B)=3.82026.02 | 39.93 | 56.59 | 48.26 | |
| BaselineFamily=Qwen 3, Size (B)=82026.02 | 39.33 | 86.99 | 63.16 | |
| BaselineFamily=Phi 4, Size (B)=3.82026.02 | 39.09 | 58.45 | 48.77 | |
| BaselineFamily=Qwen 3, Size (B)=142026.02 | 38.19 | 89.93 | 64.06 | |
| IF-FA chkpt.Family=Qwen 3, Size (B)=82026.02 | 37.95 | 84.65 | 61.3 | |
| Avg. IF chkpt.*Family=Qwen 3, Size (B)=82026.02 | 37.95 | 84.65 | 61.3 | |
| Avg. IF chkpt.Family=Qwen 3, Size (B)=142026.02 | 37.71 | 89.51 | 63.61 | |
| IF-FA chkpt.Family=Qwen 3, Size (B)=142026.02 | 36.75 | 88.55 | 62.65 | |
| BaselineFamily=Qwen 3, Size (B)=42026.02 | 35.85 | 84.71 | 60.28 | |
| IF-FA chkpt.Family=Qwen 3, Size (B)=42026.02 | 34.71 | 83.45 | 59.08 | |
| Avg. IF chkpt.*Family=Qwen 3, Size (B)=42026.02 | 34.71 | 83.45 | 59.08 | |
| IF-FA chkpt.Family=Qwen 3, Size (B)=1.72026.02 | 33.93 | 65.65 | 49.79 | |
| Avg. IF chkpt.*Family=Qwen 3, Size (B)=1.72026.02 | 33.93 | 65.65 | 49.79 | |
| BaselineFamily=Qwen 3, Size (B)=1.72026.02 | 33.87 | 70.26 | 52.07 |