Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Off-policy Evaluation research benchmarks – Page 2 · SOTA2 Research

Off-policy Evaluation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

PenDigits (UCI)

0.001MSE

6

Feb 26, 2026

0.002MSE

6

Feb 26, 2026

Sepsis Simulator

0.013RMSE

6

Feb 26, 2026

DOPE averaged (three tasks)

0.37Normalized Value Gap

6

Feb 26, 2026

MIMIC III (out-of-distribution episodes)

5.77WIS

5

May 15, 2026

D4RL HalfCheetah medium-expert

7.8RMAE

3

Feb 26, 2026

D4RL HalfCheetah medium

0.247RMAE

3

Feb 26, 2026

D4RL HalfCheetah-random

0.807RMAE

3

Feb 26, 2026

Linear Function Approximation completeness

-1Coverage

2

May 8, 2026

General Function Approximation realizability

2Coverage

2

May 8, 2026

MIMIC-III (standard random split)

67.9FQE

2

Apr 24, 2026

Real-World 2-week

92.9MDA

2

Feb 26, 2026

100MDA

2

Feb 26, 2026

D4RL Hopper-medium-expert

0.192RMAE

2

Feb 26, 2026

D4RL Hopper-medium

0.208RMAE

2

Feb 26, 2026

D4RL Walker2D-medium-expert

0.252RMAE

2

Feb 26, 2026

D4RL Walker2D-medium

0.462RMAE

2

Feb 26, 2026

D4RL Walker2D-random

0.879RMAE

2

Feb 26, 2026

Linear Function Approximation + realizability

-1Coverage

1

May 8, 2026

General Function Approximation completeness

2Coverage

1

May 8, 2026

D4RL Hopper-random

0.68RMAE

1

Feb 26, 2026

Page 2 of 2

Popular datasets

D4RL MIMIC MIMIC-III HotpotQA AlfWorld

Follow for updates

Join SOTA2 Contact