Results from SCARLET retrievals on exoTEDRF and SPARTA where C₂H₆ is included in the baseline. Top: retrievals performed on exoTEDRF dataset (data in gray, binned in purple). The best-fit model for the retrievals without (with) DMS or DMDS is shown in dark (light) purple, smoothed to R = 200. We show the impact of increasing the C₂H₆ abundance to 10⁻² in the best-fit model from the retrieval without DMS-DMDS as a dashed line. Middle: same, but using SPARTA (orange) for the NIRSpec and MIRI datasets. Bottom: 1D marginalized posterior distributions on the molecular species included in each retrieval. Colors reflect the models at the top. Black error bars and vertical lines (median ±1σ or 2σ upper limits) correspond to the combined POSEIDON results from S25 (which did not include the MIRI data).