List your company

Long-form generation on 1D-DiffTask most OOD

0.24PRRSATMD-MSP
-0.1136-0.02180.070.1618Apr 13, 2026

Evaluation results

MethodLinks
SATMD-MSPModel Backbone=Gemma-2-9B2026.040.24
SATMD-MSPModel Backbone=Llama 3.1-8B2026.040.22
SATMDModel Backbone=Gemma-2-9B2026.040.17
SATMDModel Backbone=Llama 3.1-8B2026.040.11
SATRMD-MSPModel Backbone=Llama 3.1-8B2026.040.1
HUQ-SATRMDModel Backbone=Llama 3.1-8B2026.040.09
SAPLMA (middle)Model Backbone=Llama 3.1-8B2026.040.07
SATRMDModel Backbone=Llama 3.1-8B2026.040.07
HUQ-SATMDModel Backbone=Llama 3.1-8B2026.040.07
HUQ-SATMDModel Backbone=Gemma-2-9B2026.040.07
SATRMDModel Backbone=Gemma-2-9B2026.040.06
SAPLMA (middle)Model Backbone=Gemma-2-9B2026.040.03
SAPLMA (top)Model Backbone=Gemma-2-9B2026.040.03
SATRMD-MSPModel Backbone=Gemma-2-9B2026.040.03
Uhead (2 layers)Model Backbone=Llama 3.1-8B2026.040.02
Uhead (1 layer)Model Backbone=Gemma-2-9B2026.040.02
Uhead (2 layers)Model Backbone=Gemma-2-9B2026.040.02
SAPLMA (top)Model Backbone=Llama 3.1-8B2026.040
Uhead (1 layer)Model Backbone=Llama 3.1-8B2026.04-0.01
Lookback-lensModel Backbone=Llama 3.1-8B2026.04-0.01
MSPModel Backbone=Llama 3.1-8B2026.04-0.01
Lookback-lensModel Backbone=Gemma-2-9B2026.04-0.04
HUQ-SATRMDModel Backbone=Gemma-2-9B2026.04-0.08
MSPModel Backbone=Gemma-2-9B2026.04-0.1