List your company

Short-form generation on DiffTask OOD

0.64PRRMSP
-0.3064-0.06070.1850.4307Apr 13, 2026

Evaluation results

MethodLinks
MSPModel Backbone=Gemma-2-9B2026.040.64
MSPModel Backbone=Llama 3.1-8B2026.040.57
HUQ-SATRMDModel Backbone=Gemma-2-9B2026.040.38
SAPLMA (middle)Model Backbone=Llama 3.1-8B2026.040.35
SAPLMA (middle)Model Backbone=Gemma-2-9B2026.040.3
SATRMDModel Backbone=Gemma-2-9B2026.040.27
HUQ-SATRMDModel Backbone=Llama 3.1-8B2026.040.26
HUQ-SATMDModel Backbone=Llama 3.1-8B2026.040.24
SATRMD-MSPModel Backbone=Gemma-2-9B2026.040.24
HUQ-SATMDModel Backbone=Gemma-2-9B2026.040.23
SATMDModel Backbone=Llama 3.1-8B2026.040.21
SATRMDModel Backbone=Llama 3.1-8B2026.040.21
SATRMD-MSPModel Backbone=Llama 3.1-8B2026.040.2
Uhead (1 layer)Model Backbone=Llama 3.1-8B2026.040.16
SATMD-MSPModel Backbone=Llama 3.1-8B2026.040.15
SAPLMA (top)Model Backbone=Gemma-2-9B2026.040.06
SATMDModel Backbone=Gemma-2-9B2026.040.05
Lookback-lensModel Backbone=Gemma-2-9B2026.04-0.04
SATMD-MSPModel Backbone=Gemma-2-9B2026.04-0.05
Uhead (1 layer)Model Backbone=Gemma-2-9B2026.04-0.11
Uhead (2 layers)Model Backbone=Gemma-2-9B2026.04-0.11
Lookback-lensModel Backbone=Llama 3.1-8B2026.04-0.13
SAPLMA (top)Model Backbone=Llama 3.1-8B2026.04-0.14
Uhead (2 layers)Model Backbone=Llama 3.1-8B2026.04-0.27