ResearchBenchmarksStructured report generation on SRRG-Findings (test)Follow3BLEUCoGaze1.21121.67562.142.6044Mar 27, 2026Evaluation ResultsMethodMethodLinksBLEUR-LRGPrecisionRecallF1 ScoreCoGazevariant=Llama-3Bvariant=Llama-3B2026.03321.6415.5374.8385.5678.07CoGazevariant=DistilGPT2variant=DistilGPT22026.032.820.2314.2375.8285.6178.32CheXagent2026.031.819.6515.4177.1282.5677.9RaDialog2026.031.2817.5313.8269.4870.1269.76