List your company

Mathematical Reasoning on GSM8K (Acc %, Tokens)

78.8AccuracyTL-1.5B-RL
65.59269.02172.4575.879Oct 22, 2025

Evaluation results

MethodLinks
TL-1.5B-RLBudget=100%2025.1078.8573
TL-Warmup + DiffAdaptBudget=100%2025.1078.6364
TL-Warmup + DiffAdaptBudget=72.2%2025.1078.2361
TL-Warmup + DiffAdaptBudget=50.0%2025.1077357
TL-1.5B-RLBudget=72.2%2025.1074.3510
TL-1.5B-RLBudget=50.0%2025.1072465
TL-Warmup + DiffAdaptBudget=33.3%2025.1067336
TL-1.5B-RLBudget=33.3%2025.1066.1409