Compliance evaluation on GDPR real-life contexts

92.19Accuracy

GPT-4o-mini

Updated 4d ago

Evaluation Results

Method
GPT-4o-mini 2026.04	92.19	89.29
GPT-4o 2026.04	92.19	90.02
ContextReasoner-7B-PPO 2026.04	92.19	89.33
Gemini-2.5-Flash 2026.04	92.03	89.19
GPT-4o-mini (ContextLens) 2026.04	91.87	88.25
ContextReasoner-7B-SFT 2026.04	91.71	88.36
GPT-4o-mini 2026.04	91.24	90.05
GPT-5 2026.04	91.24	88.37
o3-mini 2026.04	90.92	87.46
DeepSeek-R1 (671B) 2026.04	90.44	88.17
GPT-4o (ContextLens) 2026.04	89.31	85.55
Llama-3.1-8B-Instruct (ContextLens) 2026.04	88.85	85.13
o3-mini 2026.04	88.69	86.82
Qwen2.5-7B-Instruct 2026.04	88.37	85.36
GPT-5 2026.04	87.57	84.33
OpenThinker-7B 2026.04	87.26	81.82
Llama-3.1-8B-Instruct 2026.04	86.88	84.06
o3-mini (ContextLens) 2026.04	85.35	81.09
Qwen2.5-7B-Instruct (ContextLens) 2026.04	85.19	79.58
Gemini-2.5-Flash (ContextLens) 2026.04	83.12	76.89
GPT-5 (ContextLens) 2026.04	81.85	75.15
DeepSeek-R1 (671B) (ContextLens) 2026.04	80.89	74.52