Unverified60% confidenceBenchmarkExact time
The 2023 TruthfulQA benchmark showed that even the strongest models at the time did not exceed 60% accuracy on questions requiring precise factual judgment.
2
Sources
60%
Confidence
Long-term
Relevance
7/18/2026
First Seen
Sources
Related Claims
Unverified在MATH数据集的Level 5竞赛级难题上，前沿模型准确率仍低于60%76% similarUnverifiedA single "trusted source" prompt caused 7 out of 8 tested models to reverse between 45% and 88% of their correct answers.76% similarUnverifiedMultiple meta-analyses show that even trained personnel achieve an accuracy rate of only slightly above chance (approximately 54%–60%) when judging deception based on microexpressions.75% similarVerifiedWhen samples are severely imbalanced, Accuracy becomes misleading—a model that predicts everyone as healthy can still achieve an accuracy of over 99%.74% similarUnverifiedThe study found that the faithfulness score for counterfactual inputs was only 0.05 points lower than for factual inputs (on a 1–5 scale), a difference that is almost negligible.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/556540API
curl https://kongchang.com/api/v1/knowledge/claims/556540MCP
get_claim(id=556540)