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

Text-based Reinforcement Learning on Jericho benchmark (test)

35.8DeepHome Score

DRIFT

-0.3929.00418.427.796May 12, 2026
Updated 21d ago

Evaluation Results

MethodLinks
2026.05
35.8----------------------291.52013.8826.88252.5523.2
2026.05
6----------------------5020610255353515.3
2026.05
6----------------------7020610258250512.1
2026.05
6----------------------502065258250510.6
2026.05
3.5----------------------66.51711.45.511.757.41.25028.3
2026.05
1----------------------289.719.110.16.90030.40.53.712.6
2026.05
1----------------------207.912.117.8350.77.63409.219.2
2021.12
-2731.28.210141188101050.4555514.418118822.625.90.020.01----------
2021.12
-343514.319207.9214101050.7565517.819117.6827.330.80.060.01----------
2021.12
-33.634.51015.8246.13089.81048.251.352517.617.8117.96.927.233.1------------
2021.12
-3535181827431010105056551819118830.834.90.030----------
2021.12
-29.8401619276.7330101034.655551418118827.233.90.030.01----------
2021.12
-30.24013.821276.9330101044.7605.1917.618117.6828.235.80.030.02----------