List your company

Reading Comprehension on RACE (First-Token Accuracy)

87.3First-Token AccuracyZephyr-7B
47.98858.19468.478.606May 21, 2025

Evaluation results

MethodLinks
Zephyr-7BStrategy=Prefilling2025.0587.3
Qwen-2-7BStrategy=FTP2025.0587
Qwen-2-7BStrategy=Prompt Engineering2025.0586.7
Qwen-2-7BStrategy=Prefilling2025.0586.7
Zephyr-7BStrategy=FTP2025.0586.5
Gemma-2-9BStrategy=Prefilling2025.0586.3
Gemma-2-9BStrategy=Prompt Engineering2025.0586.2
Ministral-8BStrategy=FTP2025.0584.7
Ministral-8BStrategy=Prefilling2025.0584.7
Ministral-8BStrategy=Prompt Engineering2025.0583.8
Llama-3.1-8BStrategy=Prefilling2025.0583.2
Mistral-Nemo-12BStrategy=Prefilling2025.0583.1
Mistral-Nemo-12BStrategy=FTP2025.0582.7
Mistral-Nemo-12BStrategy=Prompt Engineering2025.0579.5
Llama-3.1-8BStrategy=Prompt Engineering2025.0578.2
Llama-3.1-8BStrategy=FTP2025.0578.1
Phi-4-14BStrategy=Prompt Engineering2025.0577.5
Phi-4-14BStrategy=Prefilling2025.0573.4
Phi-4-14BStrategy=FTP2025.0573.3
Zephyr-7BStrategy=Prompt Engineering2025.0572.3
Gemma-7BStrategy=Prefilling2025.0570.5
Gemma-7BStrategy=Prompt Engineering2025.0565.2
Gemma-7BStrategy=FTP2025.0550.4
Gemma-2-9BStrategy=FTP2025.0549.5