ResearchBenchmarksMulti-task Driving Scene Understanding Robustness on The Dolphins (Lvl. 4)Follow41.74Final ScoreNutVLM33.898435.934237.9740.0058Feb 9, 2026Evaluation ResultsMethodMethodLinksFinal ScoreNutVLM2026.0241.74Scene–CADA2026.0240.86NRP2026.0240.75Bit-Red2026.0240.6AAA2026.0240.45MS2026.0240.05JPEG2026.0239.25TVM2026.0234.2