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

Hallucination Evaluation on Object HalBench

52.7CHAIR Score (s)

baseline

0.80414.27727.7541.223Feb 4, 2026Feb 12, 2026Feb 21, 2026Mar 2, 2026Mar 10, 2026Mar 19, 2026Mar 28, 2026
Updated 18d ago

Evaluation Results

MethodLinks
2026.02
52.728----
2026.02
48.124.745.3283--
2026.02
4623----
2026.02
39.621.137.7287--
2026.02
38.419.3----
2026.02
3720.9----
2026.02
36.216.3----
2026.02
33.416.6----
2026.02
33.316.6----
2026.02
30.716----
2026.02
27.514.526.3287--
2026.03
23.810.5----
2026.02
23.713.322278--
2026.03
23.29.4----
2026.03
227.8----
2026.03
21.68----
2026.03
21.49.8----
2026.03
218----
2026.03
20.67.1----
2026.03
20.47.9----
2026.03
19.87.2----
2026.03
19.48.1----
2026.03
19.46.3----
2026.03
19.27.4----
2026.03
19.27.5----
2026.03
18.86.5----
2026.03
18.46.8----
2026.02
17.5916.7286--
2026.03
165.5----
2026.02
15.88.615.7279--
2026.03
15.85.7----
2026.03
15.45.3----
2026.02
15.38.6----
2026.02
159.2----
2026.02
14.38.5----
2026.02
7.84.2----
2026.02
6.73.6----
2026.02
5.63.2----
2026.02
5.53----
2026.02
53.1----
2026.02
4.52.4----
2026.02
4.22.5----
2026.02
42.5----
2026.02
2.91.7----
2026.02
2.91.7----
2026.02
2.81.6----
2026.01
----40.420.7
2026.01
----53.625.2
2026.01
----48.824.3
2026.01
----45.122.3
2026.01
----39.919.9
2026.01
----48.124.4
2026.01
----38.118.9
2026.01
----27.717.3
2026.01
----12.27.5
2026.01
----8.54.3
2026.01
----21.611.5
2026.01
----14.57.8
2026.01
----29.721.1
2026.01
----50.223.4
2026.01
----11.27.7