Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Question Answering on ARC Easy (Mean per-step regret)

0.096Mean Regret

TS

0.092960.113480.1340.15452Feb 23, 2026
Updated 3d ago

Evaluation Results

MethodLinks
2026.02
0.096
2026.02
0.098
2026.02
0.102
2026.02
0.103
2026.02
0.104
2026.02
0.107
2026.02
0.111
2026.02
0.115
2026.02
0.115
2026.02
0.115
2026.02
0.118
2026.02
0.121
2026.02
0.124
2026.02
0.163
2026.02
0.172