List your company

Multiple Choice Questions on TinyMMLU

86.8AccuracyPA-GRPO
52.89661.69870.579.302Mar 22, 2026

Evaluation results

MethodLinks
PA-GRPOModel=Qwen3-32B2026.0386.891.889
GRPOModel=Qwen3-32B2026.0385.291.486
PIFModel=Qwen3-32B2026.0384.39288
UniBiasModel=Qwen3-32B2026.038189.380
CalibraEvalModel=Qwen3-32B2026.0380.383.385
PriDeModel=Qwen3-32B2026.0379.984.184
BaseModel=Qwen3-32B2026.0376.58681
PIFModel=Qwen3-8B2026.0375.181.475
PA-GRPOModel=Qwen3-8B2026.0373.288.680
UniBiasModel=Qwen3-8B2026.0371.38174
PriDeModel=Qwen3-8B2026.0368.979.978
PA-GRPOModel=Llama-3.1-8B2026.0367.981.975
BaseModel=Qwen3-8B2026.0367.479.672
GRPOModel=Llama-3.1-8B2026.0367.377.973
CalibraEvalModel=Qwen3-8B2026.0363.575.568
UniBiasModel=Llama-3.1-8B2026.0361.176.565
PriDeModel=Llama-3.1-8B2026.0359.769.558
PIFModel=Llama-3.1-8B2026.0356.969.157
GRPOModel=Qwen3-8B2026.0356.683.776
CalibraEvalModel=Llama-3.1-8B2026.0355.672.162
BaseModel=Llama-3.1-8B2026.0354.268.957