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

Long-form generation on 1D-DiffTask most OOD

0.24PRR

SATMD-MSP

Updated 4d ago

Evaluation Results

Method	Links
SATMD-MSP 2026.04		0.24
SATMD-MSP 2026.04		0.22
SATMD 2026.04		0.17
SATMD 2026.04		0.11
SATRMD-MSP 2026.04		0.1
HUQ-SATRMD 2026.04		0.09
SAPLMA (middle) 2026.04		0.07
SATRMD 2026.04		0.07
HUQ-SATMD 2026.04		0.07
HUQ-SATMD 2026.04		0.07
SATRMD 2026.04		0.06
SAPLMA (middle) 2026.04		0.03
SAPLMA (top) 2026.04		0.03
SATRMD-MSP 2026.04		0.03
Uhead (2 layers) 2026.04		0.02
Uhead (1 layer) 2026.04		0.02
Uhead (2 layers) 2026.04		0.02
SAPLMA (top) 2026.04		0
Uhead (1 layer) 2026.04		-0.01
Lookback-lens 2026.04		-0.01
MSP 2026.04		-0.01
Lookback-lens 2026.04		-0.04
HUQ-SATRMD 2026.04		-0.08
MSP 2026.04		-0.1