SPHINX generates synthetic visual puzzles for benchmarking LVLMs, where GPT-5 scores 51.1% and RLVR training improves both in-domain and external visual reasoning performance.
Geometry varies in radius, sweep angle, thickness, and end-cap style
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SPHINX: A Synthetic Environment for Visual Perception and Reasoning
SPHINX generates synthetic visual puzzles for benchmarking LVLMs, where GPT-5 scores 51.1% and RLVR training improves both in-domain and external visual reasoning performance.