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

Multi-task Language Understanding on MMLU-Redux

95.2Accuracy

HieraMAS

Updated 14d ago

Evaluation Results

Method	Links
HieraMAS 2026.02		95.2	-	-	-	-	-
Self-Consistency 2026.02		94.4	-	-	-	-	-
Full-Graph 2026.02		94.4	-	-	-	-	-
CoT 2026.02		93.6	-	-	-	-	-
Self-Consistency+CoT 2026.02		93.6	-	-	-	-	-
Self-Consistency+CoT 2026.02		92.8	-	-	-	-	-
LLM-Debate 2026.02		92.8	-	-	-	-	-
Base 2026.02		92	-	-	-	-	-
LLM-Debate 2026.02		92	-	-	-	-	-
GDesigner 2026.02		92	-	-	-	-	-
Random-Graph 2026.02		91.2	-	-	-	-	-
AFlow 2026.02		91.2	-	-	-	-	-
AFlow 2026.02		91.2	-	-	-	-	-
Qwen3-30B-A3B 2026.03		90	-	-	-	-	-
Qwen3-14B 2026.03		89.9	-	-	-	-	-
Gpt-oss-20b-high 2026.03		89.8	-	-	-	-	-
CoT 2026.02		89.6	-	-	-	-	-
HieraMAS 2026.02		89.6	-	-	-	-	-
Full-Graph 2026.02		88.8	-	-	-	-	-
GDesigner 2026.02		88.8	-	-	-	-	-
MASRouter 2026.02		88.33	-	-	-	-	-
Random-Graph 2026.02		88	-	-	-	-	-
Self-Consistency 2026.02		83.2	-	-	-	-	-
Base 2026.02		82.4	-	-	-	-	-
MASRouter 2026.02		81.67	-	-	-	-	-
LCO-KLD 2026.03		76.71	-	-	-	-	-
EngGPT2-16B-A3B 2026.03		75.5	-	-	-	-	-
gemma-3-12b-it 2026.03		75.2	-	-	-	-	-
gemma-2-9b-it 2026.03		74.8	-	-	-	-	-
LCO-MSE 2026.03		74.21	-	-	-	-	-
LCO-LCH 2026.03		74.12	-	-	-	-	-
PPO 2026.03		73.93	-	-	-	-	-
GSPO 2026.03		72.72	-	-	-	-	-
GRPO 2026.03		71.35	-	-	-	-	-
DAPO 2026.03		70.37	-	-	-	-	-
Moonlight-16B-A3B-Instruct 2026.03		70.1	-	-	-	-	-
Llama-3.1-8B-Instruct 2026.03		68.4	-	-	-	-	-
NGM 2026.05		67.82	-	-	-	-	-
Qwen3-VL-2B 2026.05		67.79	-	-	-	-	-
GRPO Baseline 2026.05		62.4	0.016	-	98.4	0.991	4,973
ProFIL 2026.05		62.4	0.008	50	99.2	-	5,240
LCO-LCH 2026.03		58.71	-	-	-	-	-
LCO-KLD 2026.03		57.58	-	-	-	-	-
GRPO 2026.03		56.97	-	-	-	-	-
LCO-MSE 2026.03		56.73	-	-	-	-	-
GSPO 2026.03		56.44	-	-	-	-	-
PPO 2026.03		55.59	-	-	-	-	-
DAPO 2026.03		52.49	-	-	-	-	-