Family
System · 2024
MatPlotAgent
A plotting agent evaluated with MatPlotBench that adds execution, debugging, and rendered visual feedback.
Open the primary source ↗Task
Generate Matplotlib figures from natural-language tasks and revise them using execution and visual evidence.
Evidence
Execution and ablation results over 100 author-built benchmark tasks with limited human inspection.
What it establishes
Execution, debugging, and rendered feedback can help on a bounded plotting benchmark.
What it does not establish
An author benchmark and model judge do not establish general chart quality; some models regressed under the loop.
Related entities
Continue through the library.
- Plot2CodeA benchmark for reconstructing a scientific plot as executable plotting code from its image.
- RaivenA scientific visualization system that generates a restricted language and compiles it deterministically.
- nvAgentA system that plans and executes visualization queries over one or more databases, evaluated with VisEval.