Continual Learning on TRACE (overall)
68.1AccuracyOPR-RU
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| OPR-RUBackbone=Qwen2-7B-Instruct, Replay budget (ρ)=0.012026.05 | 68.1 | 1.97 | 56.53 | 74.73 | 65.84 | 66.15 | 88.64 | 69.75 | 79.23 | 42.97 | |
| OPR-SCBackbone=Qwen2-7B-Instruct, Replay budget (ρ)=0.012026.05 | 67.81 | 1.77 | 54.44 | 73.27 | 66.18 | 69.43 | 94.26 | 62.15 | 77.89 | 43.91 | |
| ReplayBackbone=Qwen2-7B-Instruct, Replay budget (ρ)=0.012026.05 | 65.89 | 3.37 | 53.51 | 70.31 | 60.14 | 65.18 | 93.82 | 65.43 | 77.58 | 41.12 | |
| OPR-RUBackbone=Qwen2.5-7B-Instruct, Replay budget (ρ)=0.012026.05 | 65.45 | 2.29 | 54.65 | 72.65 | 63.76 | 64.07 | 86.56 | 67.67 | 77.56 | 41.89 | |
| OPR-SCBackbone=Qwen2.5-7B-Instruct, Replay budget (ρ)=0.012026.05 | 65.11 | 2.08 | 51.86 | 70.69 | 63.6 | 66.85 | 91.68 | 59.57 | 75.31 | 41.33 | |
| SDFTBackbone=Qwen2-7B-Instruct, Replay budget (ρ)=0.012026.05 | 64.56 | 1.2 | 54.65 | 69.83 | 37.18 | 58.7 | 93.2 | 76.78 | 69.69 | 41.75 | |
| ReplayBackbone=Qwen2.5-7B-Instruct, Replay budget (ρ)=0.012026.05 | 62.57 | 4.27 | 53.51 | 69.91 | 59.74 | 64.78 | 93.42 | 65.03 | 77.08 | 40.72 | |
| OPR-RUBackbone=Llama3.1-8B-Instruct, Replay budget (ρ)=0.012026.05 | 61.89 | 2.89 | 36.9 | 71.62 | 62.53 | 64.26 | 82.14 | 41.61 | 72.26 | 42.82 | |
| OPR-SCBackbone=Llama3.1-8B-Instruct, Replay budget (ρ)=0.012026.05 | 61.52 | 2.67 | 35.53 | 70.91 | 62.36 | 65.77 | 84.5 | 39.27 | 71.7 | 42.43 | |
| SDFTBackbone=Qwen2.5-7B-Instruct, Replay budget (ρ)=0.012026.05 | 61.23 | 1 | 54.45 | 69.43 | 27.78 | 58.3 | 92.8 | 76.38 | 69.19 | 41.35 | |
| ReplayBackbone=Llama3.1-8B-Instruct, Replay budget (ρ)=0.012026.05 | 58.33 | 5.36 | 37.29 | 69 | 61.18 | 64.22 | 82.83 | 39.82 | 71.15 | 41.21 | |
| SFTBackbone=Qwen2-7B-Instruct, Replay budget (ρ)=0.012026.05 | 58.25 | 10.85 | 46.62 | 28.81 | 59.63 | 63.49 | 86.14 | 62.04 | 77.96 | 41.33 | |
| SDFTBackbone=Llama3.1-8B-Instruct, Replay budget (ρ)=0.012026.05 | 57.01 | 1.34 | 38.83 | 70.11 | 30.51 | 59.32 | 80.49 | 41.15 | 70.84 | 42.72 | |
| SFTBackbone=Qwen2.5-7B-Instruct, Replay budget (ρ)=0.012026.05 | 55.02 | 13.93 | 46.23 | 28.41 | 59.23 | 63.09 | 85.74 | 61.64 | 77.15 | 39.93 | |
| SFTBackbone=Llama3.1-8B-Instruct, Replay budget (ρ)=0.012026.05 | 50.78 | 17.07 | 23.26 | 43.69 | 60.59 | 64.27 | 72.83 | 28.09 | 72.15 | 41.32 |