Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Generative Hallucination on AMBER Generative

70.4Coverage (%)

Qwen2.5-VL 3B + SFT

43.46450.45757.4564.443Mar 1, 2025Apr 28, 2025Jun 26, 2025Aug 24, 2025Oct 22, 2025Dec 20, 2025Feb 17, 2026
Updated 11d ago

Evaluation Results

MethodLinks
2025.10
70.4-48.83.96.9
2025.10
69.9-55.75.87.6
2025.10
68.5-44.64.26.7
2025.10
67.7-44.33.46.6
2025.03
67.14.630.72.6-
2025.10
67.1-30.72.64.6
2025.10
64.9-32.42.15.4
2025.10
64.3-28.91.74.9
2025.10
63.6-27.51.64.8
2025.10
63.2-23.91.24.2
2025.10
57.2-23.61.35.6
2025.03
536.632.23.4-
2026.01
52.86.934.23.7-
2026.01
52.47.134.73.9-
2026.02
51.87.131.43.4-
2026.02
51.67.6364-
2026.01
517.836.44.2-
2026.02
517.836.44.2-
2025.10
51-36.44.27.8
2026.01
50.87.436.14-
2026.02
50.87.536.24.1-
2026.02
49.97.433.23.7-
2025.03
49.86.730.93.3-
2026.02
49.67.3323.5-
2025.10
49.4-23.61.95.5
2025.03
49.24.823.41.2-
2025.03
49.15.122.72-
2025.03
48.96261.5-
2025.03
48.56.122.21.3-
2025.03
47.77.629.92.2-
2025.03
47.26.927.52.2-
2025.03
46.76.7282.6-
2025.03
46.66.325.62-
2025.03
46.56.727.82-
2025.03
46.48.431.12.6-
2025.03
44.58312.2-