RLAIF-V: Open-Source AI Feedback Leads to Super GPT-4V Trustworthiness
Open paper
Benchmark
Model
Rank
Results
image-captioning-on-object-halbench
RLAIF-V 7B
#2
chair_i: 4.3
chair_s: 8.5
image-captioning-on-object-halbench
RLAIF-V 12B
#3
chair_i: 1.8
chair_s: 3.3
Rank counts only results with a code link.