Process Reward Models (PRMs) trained on step-level error labels automatically annotated by formal verification tools.
Ryo Kamoi
ryokamoi
AI & ML interests
NLP
Organizations
Dataset for evaluating the visual perception capabilities of LVLMs.
-
VisOnlyQA: Large Vision Language Models Still Struggle with Visual Perception of Geometric Information
Paper • 2412.00947 • Published • 8 -
ryokamoi/VisOnlyQA_Eval_Real_v1.1
Viewer • Updated • 900 • 75 -
ryokamoi/VisOnlyQA_Eval_Synthetic
Viewer • Updated • 700 • 85 • 2 -
ryokamoi/VisOnlyQA_Train
Viewer • Updated • 70k • 204 • 2
Process Reward Models (PRMs) trained on step-level error labels automatically annotated by formal verification tools.
Dataset for evaluating the visual perception capabilities of LVLMs.
-
VisOnlyQA: Large Vision Language Models Still Struggle with Visual Perception of Geometric Information
Paper • 2412.00947 • Published • 8 -
ryokamoi/VisOnlyQA_Eval_Real_v1.1
Viewer • Updated • 900 • 75 -
ryokamoi/VisOnlyQA_Eval_Synthetic
Viewer • Updated • 700 • 85 • 2 -
ryokamoi/VisOnlyQA_Train
Viewer • Updated • 70k • 204 • 2
Benchmark for error detection in LLM responses