SOTAVerified

Factual Inconsistency Detection in Chart Captioning

Detect factual inconsistency between charts and captions.

Papers

No papers found.

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1Bard (before Gemini)Kendall's Tau-c0.29Unverified
2GPT-4VKendall's Tau-c0.22Unverified
3ChartVEKendall's Tau-c0.22Unverified
4LLaVA-1.5-13BKendall's Tau-c0.21Unverified
5DePlot + GPT-4Kendall's Tau-c0.11Unverified
#ModelMetricClaimedVerifiedStatus
1GPT-4VKendall's Tau-c0.21Unverified
2DePlot + GPT-4Kendall's Tau-c0.12Unverified
3BardKendall's Tau-c0.11Unverified
4ChartVEKendall's Tau-c0.09Unverified
5LLaVA-1.5-13BKendall's Tau-c0.06Unverified
#ModelMetricClaimedVerifiedStatus
1ChartVEKendall's Tau-c0.18Unverified
2GPT-4VKendall's Tau-c0.16Unverified
3DePlot + GPT-4Kendall's Tau-c0.13Unverified
4LLaVA-1.5-13BKendall's Tau-c0Unverified
5Bard (before Gemini)Kendall's Tau-c-0.01Unverified
#ModelMetricClaimedVerifiedStatus
1ChartVEKendall's Tau-c0.18Unverified