A More Results

The overall performance in MM-NIAH is shown in Tab. 2, which is obtained by averaging the performance across the six tasks in We also provide the performance of each task in Tab. A.1 More findings In addition to the findings discussed in Section 4.2, we provide more findings here. Placing questions before context does NOT improve model performance. Therefore, we do not provide quantitative results but qualitatively analyzed this issue. The long context understanding ability of Gemini-1.5 is not perfect.