.4
61.6
60.0
64.7
0.1
4.0K
46s
71.8
60.5
59.9
64.1
0.2
3.5K
23s
70.3
60.4
57.8
62.8
0.3
2.4K
18s
73.4
60.7
61.2
65.1
0.4
1.6K
13s
70.1
60.6
57.5
62.7
0.5
1.2K
12s
68.9
58.2
55.3
60.8
1.0
0.0K
10s
66.5
56.4
53.7
58.9
Table 6. Comparison with other RAG methods across different backbones on Video-MME (Fu et al., 2025).
Model
Params
Frames
Short
Medium
Long
Overall
X=Video-LLaVA (Lin et al., 2024a)
7B
8
44.6
38.3
35.8
39.6
X+Video-RAG (Luo et al., 2024)
7B
8
49.5
43.0
42.5
45.0
X+TV-RAG (Cao et al., 2025)
7B
8
49.7
44.3
42.6
45.5
X+EC-RAG
7B
8
50.1
44.6
43.4
46.0
X=LongVA (Zhang et al., 2024b)
7B
32
60.9
49.3
44.0
51.4
X+Video-RAG (Luo et al., 2024)
7B
32
65.4
59.1
55.7
60.1
X+TV-RAG (Cao et al., 2025)
7B
32
66.2
62.1
58.7
61.7
X+EC-RAG
7B
32
67.1
62.5
58.9
62.8
Appendix B Task-Type Breakdown
Figure 5. Accuracy of LLaVA-Video-7B (Zhang et al., 2024e) with and without EC-RAG under different frame budgets on Video-MME (Fu et al., 2025). EC-RAG provides consistent gains at all budgets, with the largest improvement at 16 frames.Bar chart comparing baseline and EC-RAG accuracy across 8, 16, and 32 frames on short, medium, and long videos.
Table 7. Per-task-type accuracy (%) on Video-MME (Fu et al., 2025) with LLaVA-Video-7B (Zhang et al., 2024e) as the backbone. Gain denotes the absolute improvement over the baseline. Best results per frame setting are in bold.
Method
Frames
TmpR
ActR
ActRec
AttrP
SpaP
SpaR
TmpP
InfoS
OCR
ObjR
ObjRec
Count
AVG
Gain
LLaVA-Video (Zhang et al., 2024e)
8
41.8
47.4
55.6
69.4
59.3
71.4
56.4
70.0
48.9
51.1
58.5
37.7
54.6
–
LLaVA-Video + EC-RAG
8
46.9
49.8
55.9
73.9
59.6
80.4
60.0
79.6
56.8
60.6
63.6
40.3
60.2
+5.6
LLaVA-Video (Zhang et al., 2024e)
16
45.2
51.2
57.5
68.0
57.4
73.2
69.1
74.0
50.4
55.4
64.1
41.8
58.0
–
LLaVA-Video + EC-RAG
16
51.4
64.9
62.6
78.4
61.1
80.4
69.6
83.9
59.0
66.1
67.2
42.3
65.8
+7.8
Table 7 presents a fine-grained breakdown of accuracy across all 12 task types on Video-MME under 8-frame and 16-frame settings. With 8 frames, EC-RAG improves the average by +5.6 points (54.6%→\to60.2%), with the largest gains on Information Synopsis (+9.6), Spatial Reasoning (+9.0), and OCR (+7.9)—tasks that heavily rely on textual or spatial cues where event-chain evidence provides direct support. At 16 frames, the average gain further widens to +7.8 points (58.0%→\to65.8%), with Action Reasoning seeing the most dramatic improvement (+13.7) followed by Attribute Perception (+10.4) and Information Synopsis (+9.9). EC-RAG achieves the best score on 10 out of 12 task types at 8 frames and 12 out of 12 at 16 frames, demonstrating that the event-chain retrieval complements visual features comprehensively across diverse question types rather than benefiting only a narrow subset.
Figure 6. Qualitative example 1: a history documentary where the question requires reasoning about the temporal order of events after a key battle. The baseline confuses visually similar city depictions, while EC-RAG traces the event chain through ASR narration to identify the correct subsequent event.Qualitative comparison between baseline and EC-RAG on a history documentary temporal reasoning question.
Figure 7. Qualitative example 2: a science documentary where the question asks about the introduction order of four topics. The baseline is misled by recurring visual elements across segments, while EC-RAG recovers the correct topic sequence from the structured event chain.Qualitative comparison between baseline and EC-RAG on a science documentary topic ordering question.
Appendix C More Qualitative Results
We present additional qualitative results of LLaVA-Video-7B (Zhang et al., 2024e) with EC-RAG on two representative examples from Video-MME (Fu et al., 2025) in Figures 6 and 7. Both cases involve temporal reasoning questions from long documentary videos, where the baseline model fails due to the inherent ambiguity of sparsely sampled frames. As illustrated, by constructing temporal event chains from multimodal evidence—particular