List your company
ResearchFormerly WizWand

Instruction Following on MQUAKE

82.5AccuracyLLAMA3-8B-INSTRUCT
52.44460.24768.0575.853Dec 18, 2024
Updated 8mo ago

Evaluation results

MethodLinks
LLAMA3-8B-INSTRUCTAlignment=Context-DPO, Few-shot=5-shot2024.1282.5
MISTRAL-7B-INSTRUCTAlignment=Context-DPO, Few-shot=5-shot2024.1282.3
QWEN2-7B-INSTRUCTAlignment=Context-DPO, Few-shot=5-shot2024.1281.8
LLAMA2-7B-CHATAlignment=Context-DPO, Few-shot=5-shot2024.1280.2
LLAMA3-8B-INSTRUCTAlignment=Context-DPO, Few-shot=3-shot2024.1279.7
QWEN2-7B-INSTRUCTAlignment=Context-DPO, Few-shot=3-shot2024.1278.6
MISTRAL-7B-INSTRUCTAlignment=Context-DPO, Few-shot=3-shot2024.1277.8
LLAMA2-7B-CHATAlignment=Context-DPO, Few-shot=3-shot2024.1276.5
LLAMA2-7B-CHATAlignment=Base, Few-shot=5-shot2024.1274.5
QWEN2-7B-INSTRUCTAlignment=Context-DPO, Few-shot=1-shot2024.1274.4
MISTRAL-7B-INSTRUCTAlignment=Context-DPO, Few-shot=1-shot2024.1274.2
LLAMA2-7B-CHATAlignment=Context-DPO, Few-shot=1-shot2024.1272.6
LLAMA3-8B-INSTRUCTAlignment=Context-DPO, Few-shot=1-shot2024.1271.3
QWEN2-7B-INSTRUCTAlignment=Base, Few-shot=5-shot2024.1271.3
LLAMA2-7B-CHATAlignment=Base, Few-shot=3-shot2024.1268.1
LLAMA3-8B-INSTRUCTAlignment=Base, Few-shot=5-shot2024.1267.9
LLAMA2-7B-CHATAlignment=Base, Few-shot=1-shot2024.1267.3
QWEN2-7B-INSTRUCTAlignment=Base, Few-shot=3-shot2024.1264.2
LLAMA3-8B-INSTRUCTAlignment=Base, Few-shot=3-shot2024.1263.8
QWEN2-7B-INSTRUCTAlignment=Base, Few-shot=1-shot2024.1261
MISTRAL-7B-INSTRUCTAlignment=Base, Few-shot=5-shot2024.1258.4
LLAMA3-8B-INSTRUCTAlignment=Base, Few-shot=1-shot2024.1256.8
MISTRAL-7B-INSTRUCTAlignment=Base, Few-shot=3-shot2024.1255.9
MISTRAL-7B-INSTRUCTAlignment=Base, Few-shot=1-shot2024.1253.6