List your company

Continual Instruction Following on ALFRED

69.9Success Rate (SR)ExRAP
-2.79616.07734.9553.823Oct 4, 2021May 31, 2022Jan 26, 2023Sep 22, 2023May 19, 2024Jan 13, 2025Sep 10, 2025

Evaluation results

MethodLinks
ExRAPNon-stationarity=Low2025.0969.95.94
ExRAPNon-stationarity=Medium2025.09644.82
ExRAPNon-stationarity=High2025.0959.114.42
LLM-PlannerNon-stationarity=Low2025.0958.447.28
LLM-PlannerNon-stationarity=Medium2025.0951.87.28
ProgPromptNon-stationarity=Low2025.0947.159.81
SayCanNon-stationarity=Low2025.0945.678.25
SayCanNon-stationarity=Medium2025.0941.818.39
(SL)^3 + plannerData percentage=10%, Supervision level=Extra information, Modular components=Planner, Extra info details=Includes ground-truth training segmentations/alignments and ground-truth test-time plans2021.1040.4—
SayCanNon-stationarity=High2025.0935.797.42
LLM-PlannerNon-stationarity=High2025.0935.766.65
ProgPromptNon-stationarity=Medium2025.0935.627.22
FILMSupervision level=Extra information, Extra info details=Includes ground-truth training segmentations/alignments and ground-truth test-time plans2021.1020.1—
ProgPromptNon-stationarity=High2025.0919.977.52
ZSPNon-stationarity=Low2025.0918.2217.24
HLSMSupervision level=Goal + partial plan sup., Reproduction=Author's reproduction, Test-time info=Goals only2021.1017.2—
(SL)^3 + HLSMData percentage=10%, Supervision level=Goal + partial plan sup., Modular components=HLSM, Test-time info=Goals only2021.1016.1—
ZSPNon-stationarity=Medium2025.0914.6720.83
ABPSupervision level=Extra information, Extra info details=Includes ground-truth training segmentations/alignments and ground-truth test-time plans2021.1012.6—
HiTUTSupervision level=Extra information, Extra info details=Includes ground-truth training segmentations/alignments and ground-truth test-time plans2021.1011.1—
ZSPNon-stationarity=High2025.099.5622.53
ETSupervision level=Extra information, Extra info details=Includes ground-truth training segmentations/alignments and ground-truth test-time plans2021.107.3—
EmBERTSupervision level=Extra information, Extra info details=Includes ground-truth training segmentations/alignments and ground-truth test-time plans2021.105.7—
MOCASupervision level=Extra information, Extra info details=Includes ground-truth training segmentations/alignments and ground-truth test-time plans2021.105.4—
S+Supervision level=Extra information, Extra info details=Includes ground-truth training segmentations/alignments and ground-truth test-time plans2021.100.1—
(SL)^3Data percentage=10%, Supervision level=Goal + partial plan sup., Test-time info=Goals only2021.100—
seq2seqSupervision level=Goal + partial plan sup., Test-time info=Goals only2021.100—
seq2seq2seqSupervision level=Goal + partial plan sup., Test-time info=Goals only2021.100—