SEND and CDPK benchmark results correlate highly (r = 0.94), but higher-performing models do relatively better on general pedagogy than SEND¶

Claim · All claims
Evidence · 1 study · design r3 · q2 · i3 large
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person · recheck by 2028: tool and model accuracy changes quickly

Subclaims¶

q2 i3 SEND and CDPK accuracies across models are highly correlated (r = 0.94), yet models above 60% accuracy tended to do better on general pedagogy than SEND, while for lower-performing models this was reversed. → Lelièvre 2025

Evidence¶

Lelièvre 2025¶

Lelièvre, M., Waldock, A., Liu, M., Valdes Aspillaga, N., Mackintosh, A., Ogando Portela, M. J., Lee, J., Atherton, P., Ince, R. A. A., & Garrod, O. G. B. (2025). Benchmarking the Pedagogical Knowledge of Large Language Models. https://arxiv.org/abs/2506.18710 doi:10.48550/arxiv.2506.18710

q2 · i3 · design · r3

Benchmark evaluation of models on the 220 tested SEND questions compared with their CDPK results, shown in Figure 12. The article prints the correlation coefficient r = 0.94 between the two benchmarks' accuracies and reports the performance reversal across model tiers.

"While the results of the two benchmarks were overall quite highly correlated (𝑟 = 0.94), it's interesting that higher performing models (above60% accuracy) tended to do better on general pedagogy than SEND, while for lower performing models, this was reversed."

Discussion¶

Related Claims¶

• The benchmark procedure is stable under stochastic decoding, with a maximum accuracy standard deviation of 0.43% over 20 repeated runs — related
• Estimated human baseline on the CDPK benchmark is approximately 50%, based on aggregate Chilean teacher exam results — related
• Lower-performing LLMs show higher variance across pedagogy subject categories and peak in Technology and General categories — related