AI Hallucination vs. Accuracy
AI Hallucination vs. Accuracy
AI models are advancing quickly and there are more and more players in this field. However, a common challenge with all AI models is hallucination. Though I like how AI model makes coding easier, I still cannot trust the code it generated 100%. This week, the dataset I am visualizing is the hallucination vs. accuracy of various AI models. The original visualization was posted on Visual Capitalist.
My Visualization
This is a scatter plot showing accuracy vs. halluciation. Models from the same company share the same color.
Please note that all the visualizations are designed for desktop view, so it is recommended to view them on a desktop device.
Insights
- The two Claude models (Sonnet and Opus) has the lowest(best) halluciantion index and relatively good accuracy;
- Meanwhile, gpt-oss models hallucinate the most with poor accuracy.
Follow this link to find more weekly vizzes :)
