Commonsense Reasoning Benchmark benchmark leaderboard · SOTA2 Research

Commonsense Reasoning Benchmark

75.91BoolQ AccuracyStelLA

Evaluation Results

Method	Links
StelLA2025.10		75.91	89.86	81.68	96.41	87.82	91.98	82.34	87.8	86.72
DoRA2025.10		75.38	88.01	79.94	95.35	86.29	90.54	79.69	86.07	85.16
ScaledAdamW2025.10		75.24	88.57	80.21	95.81	85.11	91.09	80.55	86.6	85.4
LoRA2025.10		75.16	88.14	80.18	95.41	86.74	90.84	78.7	87	85.27
MoSLoRA2025.10		74.88	88.43	80.31	95.5	86.26	90	79.86	85.8	85.13
PiSSA2025.10		74.67	88.12	80.5	94.98	85.22	90.15	78.87	85.6	84.76
OLoRA2025.10		74.41	87.68	79.55	94.79	85.4	90.04	78.24	85	84.39
StelLA2025.10		73.62	84.87	80.64	91.44	84.5	86.43	72.84	84.33	82.33
LoRA2025.10		73.2	85.8	81.8	94.9	86	74.7	88.7	86	83.9
LoRA-DA2025.10		73.2	86	82.4	95.2	87.1	75.7	88.7	86.4	84.3
ChatGPT2025.10		73.1	85.4	68.5	78.5	66.1	89.8	79.9	74.8	77
MiLoRA2025.10		73.1	85.3	81.8	95.1	86.3	75.3	88.6	86.8	84
TriLoRA2025.10		73.09	86.64	78.64	93.4	82.88	87.76	75.26	84.3	82.74
LoRA-One2025.10		72.8	85.3	81.9	95.2	85.6	74.9	88.8	86.4	83.9
DoRA2025.10		72.67	83.48	79.82	90.82	83.58	85.16	71.27	81.2	81
PiSSA2025.10		72.5	85.2	81.9	94.2	86.7	73.7	87.1	87	83.5
ScaledAdamW2025.10		72.2	83.86	79.67	90.8	82.43	85.55	70.59	81.93	80.88
LoRA2025.10		72.02	83.46	79.87	90.44	82.69	84.83	71.19	81.53	80.76
MoSLoRA2025.10		71.54	83.84	79.6	90.5	83.19	84.4	69.96	80.47	80.44
TriLoRA2025.10		71.23	80.96	78.33	80.91	77.59	81.76	66.69	79.8	77.16
PiSSA2025.10		71.16	83.89	79.19	91	82.87	85.09	69.48	83.93	80.83
OLoRA2025.10		71.11	82.7	78.64	89.41	81.48	83.58	68.17	80.2	79.41