The experiment evaluates LLM-agent replanning under a transient actuator degradation injection. Fault onset is randomized between environment steps 40 and 60.
Expected evaluation record
Metric
Recorded value
Degradation-area
11.91 ± 4.23 (corrected; mean ± 95% CI, n=72)
Safety-violation excess
60.8 ± 21.7 (weighted, control-subtracted; mean ± 95% CI, n=72)
Fraction of time healthy
0.581 ± 0.026 (mean ± 95% CI, n=72)
Relapse count
0.56 ± 0.16 (mean ± 95% CI, n=72)
False-recovery rate
0.832 ± 0.034 (mean ± 95% CI, n=72)
Intervention recall
0.413 ± 0.072 (mean ± 95% CI, n=72)
Values are means with 95% confidence intervals across every target × duration-fixed combination shown on this page (see graphs below), 8 seeds each. See the Results page for the cross-mechanism comparison.
Experimental trajectory graphs
One representative seed (seed 0 of 8) per target × mode combination from the full sweep. See the Results page for the multi-seed aggregate comparison across mechanisms. The graphs for every seed are archived at doi:10.5281/zenodo.23304785 (files available from 10 November 2026).
Sorting Agent Target (restrict mode)
LLM-agent replanning - Sorting Agent Target (restrict mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Pressing Agent Target (restrict mode)
LLM-agent replanning - Pressing Agent Target (restrict mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Both Agents Target (restrict mode)
LLM-agent replanning - Both Agents Target (restrict mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Sorting Agent Target (slip mode)
LLM-agent replanning - Sorting Agent Target (slip mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Pressing Agent Target (slip mode)
LLM-agent replanning - Pressing Agent Target (slip mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Both Agents Target (slip mode)
LLM-agent replanning - Both Agents Target (slip mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Sorting Agent Target (stuck mode)
LLM-agent replanning - Sorting Agent Target (stuck mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Pressing Agent Target (stuck mode)
LLM-agent replanning - Pressing Agent Target (stuck mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Both Agents Target (stuck mode)
LLM-agent replanning - Both Agents Target (stuck mode) - evaluation trajectory under transient actuator degradation (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep.
Notes
This page is intentionally separated from the paper-facing results so that full-resolution plots, additional seeds, and diagnostic figures can be documented without overloading the publication.
