List your company

Mathematical Reasoning on GSM (test)

65.4AccuracyTeacher
-2.61615.04232.750.358Oct 9, 2025Nov 3, 2025Nov 29, 2025Dec 25, 2025Jan 20, 2026Feb 15, 2026Mar 13, 2026

Evaluation results

MethodLinks
TeacherTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1065.4——
TeacherTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1059.6——
TeacherTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1036.9——
AdaSwitchTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1036.3——
SWITCHTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1035.2——
AdaSwitchTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1035——
ImitKDTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1034.6——
On-policyTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1034.1——
Off-policyTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1033.5——
SWITCHTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1033.2——
SKDTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1033——
On-policyTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1032.7——
SKDTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1032.4——
SFTTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1031——
Off-policyTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1030.6——
ImitKDTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1030.6——
SeqKDTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.1030.4——
SFTTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1029——
SeqKDTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.1028.2——
AdaSwitchTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1026.8——
On-policyTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1026.2——
SKDTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1026——
ImitKDTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1025.3——
SWITCHTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1024.8——
Off-policyTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1022.6——
SeqKDTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1021.8——
SFTTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.1018.9——
Alpaca-GPT4 + NAIT (7-11)System ID=12, Capability Features=All, Data sampling ratio=10%2026.0316.5337.23.24
Alpaca-GPT4 + NAIT (GSM)System ID=08, Capability Features=Individual, Data sampling ratio=10%2026.031637.74.65
LIMASystem ID=02, Fine-tuning protocol=Full Fine-tuning, Data sampling ratio=10%2026.0315.7636.170.39
Alpaca-GPT4 + NAIT (MMLU)System ID=07, Capability Features=Individual, Data sampling ratio=10%2026.0315.6837.183.2
Alpaca-GPT4 + NAIT (CodeX)System ID=09, Capability Features=Individual, Data sampling ratio=10%2026.0315.5337.062.88
Alpaca-GPT4 + SelectITSystem ID=05, Fine-tuning protocol=Full Fine-tuning, Data sampling ratio=10%2026.0315.437.163.15
Alpaca-GPT4System ID=01, Fine-tuning protocol=Full Fine-tuning, Data sampling ratio=10%2026.0314.6336.03—
Alpaca-GPT4 + Q2QSystem ID=04, Fine-tuning protocol=Full Fine-tuning, Data sampling ratio=10%2026.0314.535.68-0.98
Alpaca-GPT4 + Random BaselineSystem ID=06, Fine-tuning protocol=Full Fine-tuning, Data sampling ratio=10%2026.0314.1335.69-0.94
Alpaca-GPT4 + NAIT (TydiQA)System ID=10, Capability Features=Individual, Data sampling ratio=10%2026.0313.836.711.89
Alpaca-GPT4 + AlpaGasusSystem ID=03, Fine-tuning protocol=Full Fine-tuning, Data sampling ratio=10%2026.0313.3435.18-2.34
Alpaca-GPT4 + NAIT (BBH)System ID=11, Capability Features=Individual, Data sampling ratio=10%2026.0313.3436.471.23
StudentTeacher and Student Model Pair=Llama-3.1-3b to Llama-3.1-1b2025.100——
StudentTeacher and Student Model Pair=Qwen-2.5-3b to Qwen-2.5-0.5b2025.100——
StudentTeacher and Student Model Pair=Gemma-7b to Gemma-2b2025.100——