Calibration and Measurement on 14 social science constructs (Sample Dataset) benchmark leaderboard · SOTA2 Research

Calibration and Measurement on 14 social science constructs (Sample Dataset)

0.186T-ECE EpsilonQwen2.5-7B

Evaluation Results

Method	Links
Qwen2.5-7B2026.05		0.186	0.265	0.409
Gemma-2-9B2026.05		0.216	0.256	0.429
Ministral-8B2026.05		0.23	0.254	0.254
DeepSeek-R1-7B2026.05		0.234	0.231	0.26
GPT-5-nano2026.05		0.354	0.359	0.451
DeepSeek-V3.22026.05		0.366	0.382	0.467
Ministral-8B2026.05		0.374	0.327	0.254
DeepSeek-R1-7B2026.05		0.378	0.367	0.283
Qwen2.5-7B2026.05		0.417	0.411	0.402
Qwen2.5-7B2026.05		0.426	0.414	0.455
DeepSeek-V3.22026.05		0.431	0.413	0.501
Ministral-8B2026.05		0.455	0.383	0.291
GPT-5-mini2026.05		0.466	0.438	0.574
Gemma-2-9B2026.05		0.495	0.454	0.429
Gemma-2-9B2026.05		0.498	0.476	0.418
Ministral-8B2026.05		0.511	0.495	0.465
Qwen3.5-Flash2026.05		0.528	0.503	0.544
DeepSeek-R1-7B2026.05		0.538	0.479	0.236
Gemma-2-9B2026.05		0.565	0.543	0.443
GPT-5-nano2026.05		0.568	0.558	0.478
Qwen3.5-Flash2026.05		0.604	0.588	0.548
Qwen2.5-7B2026.05		0.626	0.603	0.409
GPT-5-mini2026.05		0.654	0.648	0.559
DeepSeek-R1-7B2026.05		0.723	0.691	0.26