Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Multi-task Language Reasoning on RLU Suite (GSM8K, ARC-C, MEDQA, BOOLQ, COLA)

90.67GSM8K Accuracy

Hard-Routed MoR-LoRA

74.279678.534882.7987.0452Jun 30, 2026
Updated 25d ago

Evaluation Results

MethodLinks
2026.06
90.6789.2558.1388.183.4181.95
2026.06
90.688.6556.285.5481.7880.55
2026.06
87.5783.6148.7884.881.5977.27
2026.06
86.8883.147.6883.4381.376.48
2026.06
86.0583.746.580.9480.4475.53
2026.06
74.9188.9957.58383.1377.51